You are on page 1of 254

Metaphor across Time and Conceptual Space

Cognitive Linguistic Studies

in Cultural Contexts (CLSCC)
This book series aims at publishing high-quality research on the relationship between
language, culture, and cognition from the theoretical perspective of Cognitive
Linguistics. It especially welcomes studies that treat language as an integral part
of culture and cognition, that enhance the understanding of culture and cognition
through systematic analysis of language qualitative and/or quantitative, synchronic
and/or diachronic and that demonstrate how language as a subsystem of culture
transformatively interacts with cognition and how cognition at a cultural level is
manifested in language.
For an overview of all books published in this series, please see

Ning Yu and Farzad Sharifian
Pennsylvania State University / Monash University

Editorial Board
Antonio Barcelona Roslyn M. Frank Fiona MacArthur
Universidad de Crdoba University of Iowa Universidad de Extremadura
Erich A. Berendt Raymond W. Gibbs, Jr. Todd Oakley
Assumption University, University of California, Santa Case Western Reserve
Bangkok Cruz University
Alan Cienki Masako K. Hiraga Chris Sinha
VU University Amsterdam Rikkyo University University of Portsmouth
Alice Deignan Zoltn Kvecses Gerard J. Steen
University of Leeds Etvs Lornd University VU University Amsterdam
Vyvyan Evans Graham Low Hans-Georg Wolf
Bangor University University of York Potsdam University
Charles Forceville Zouhair Maalej
University of Amsterdam King Saud University

Volume 3
Metaphor across Time and Conceptual Space. The interplay of embodiment and
cultural models
by James J. Mischler, III
Metaphor across Time
and Conceptual Space
The interplay of embodiment and cultural models

James J. Mischler, III

Northwestern State University of Louisiana

John Benjamins Publishing Company

The paper used in this publication meets the minimum requirements of

theAmerican National Standard for Information Sciences Permanence

of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data

Mischler, James J., III.

Metaphor across time and conceptual space : the interplay of embodiment and cultural
models / James J. Mischler, III.
p. cm. (Cognitive Linguistic Studies in Cultural Contexts, issn 1879-8047 ; v. 3)
Includes bibliographical references and index.
1. Space and time in language. 2. Metaphor. 3. Language and culture. 4. Cognitive gram-
mar. 5. Psycholinguistics. I. Title.
P37.5.S65M57 2013
401--dc23 2013012311
isbn 978 90 272 0406 6 (Hb ; alk. paper)
isbn 978 90 272 7180 8 (Eb)

2013 John Benjamins B.V.

No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any
other means, without written permission from the publisher.
John Benjamins Publishing Co. P.O. Box 36224 1020 me Amsterdam The Netherlands
John Benjamins North America P.O. Box 27519 Philadelphia pa 19118-0519 usa
To my parents
Table of contents

Tables and figures xv

part i. Theoretical foundations

chapter 1
The Cognition-Culture interface 3
Introduction 3
Key theoretical constructs 3
Cognitive-Functionalism 3
Conceptualization 6
Non-autonomous knowledge 7
Culture and cultural models 7
Cultural models as conceptual systems 9
Cultural models as conceptual systems: An example 11
Cultural models and syntax: Another example 13
Cultural models provide perspective on a scene 14
The implications of cultural models as cognitive constructs 14
Summary 15
Usage-based theory of language 15
Usage-based models and empirical research 16
Is conceptualization dynamic? 17
Form-meaning pair 19
The semiotic triangle: Form, meaning, and community
common ground 19
Theories of shared cultural knowledge 20
Understanding ambiguous utterances 21
Intersubjectivity revisited 22
Summary 22
Linguistic metaphor 23
Methodological issues 23
The role of non-linguistic data 24
Implications for the current studies 25
Non-linguistic data in conceptual metaphor research 25
Metaphor across Time and Conceptual Space

Summary 26
The use of text corpora 27
Introspection as a data collection strategy 27
Introspection and the problem of context 27
Text corpora as a data collection strategy 28
Corpus research in Cognitive-Functionalism 29
Corpus methods in perspective 30
Summary 33
Corpus-assisted discourse studies (CADS) 34
Chapter summary 36
Plan of the volume 36

chapter 2
Diachronic aspects of synchronic concepts 39
Introduction 39
Introduction to conceptual metaphor theory 39
Perspective on a scene via cultural models 41
Analysis of the CM of anger 42
The bodily experience of anger 42
Sub-variations of the CM of anger 43
The fluid CM 44
Elaborations of the fluid CM 45
A research gap 46
The experiential scene 47
The anger prototype scenario 47
Cultural models as bases in conceptualization 48
Atypical cases of anger 49
controlled response over time 49
intense response over time 50
Conclusions 52
Classification of the non-prototypical cases 52
control in the non-prototypical cases 52
Theory as a test of observed language patterns 53
The proposal: The blood and spleen metaphors 54
Synchronic studies of CM and culture 54
Conceptual metaphors as universals 54
Matsuki (1995) 54
Yu (1995) 55
Conceptual metaphors as cultural models 56
Maalej (2004) 56
Table of contents

Kvecses (2010a) 57
Conclusion: embodiment and cultural models as equal partners 58
Frequency of use and cultural models 59
Chapter summary and conclusions 60
Filling the research gap 61

chapter 3
Metaphor across historical time 63
Introduction 63
Two types of historical study 64
Synchronic-historical research 64
Synchronic-historical research: An example 64
A synchronic-historical study of CM and culture 65
Summary 67
Diachronic studies of conceptual metaphor 67
Diachronic studies: Single factor designs 68
Gevaert (2002) 68
Geeraerts and Gevaert (2008) 70
Koivisto-Alanko and Tissari (2006) 72
Diachronic designs: Multiple factor studies 74
Trim (2011) 74
Geeraerts, Gevaert, and Speelman (2011) 75
Summary 77
The research gap 78
Theoretical implications 79
Methodological implications: The role of frequency statistics 79
Motivations for the current studies 80

part ii. A macro-study of human emotion in cultural context,

A.D. 15001990

chapter 4
Research questions and methodology 85
Introduction 85
Research questions 85
The ancillary study of historical non-linguistic data 86
Data collection 86
The four types of data 86
Data analysis 90
Metaphor across Time and Conceptual Space

The main study of diachronic linguistic expressions of anger 90

Materials 90
Data 91
Implications for data identification and collection 92
Selecting the search keywords 93
The role of context 93
Limitations of the keyword search procedure 93
Data collection 94
Data analysis 94
An implication of the method: The value of mixed research designs 98
Chapter summary 99

chapter 5
Results of the ancillary study of non-linguistic data 103
Introduction 103
The ancillary study of the four humors cultural model 104
General principles of the four humors 104
The self-care focus 104
The macrocosm/microcosm cultural model 105
Implications for the four humors model 106
Summary 108
The ancillary study of non-linguistic data: Results 108
The unified model: A historical composite view 109
Basic principles of the unified model 109
The macrocosm principle 110
The microcosm principles 110
The concept of balance in the four humors 115
Five scientific advances in human physiology, A.D. 15001990 116
15001599: Scientific anatomy (Andreas Vesalius, 1543) 117
Prior research 117
Vesalius work and influence 118
Evidence for the unified model in 16th century lay practice 118
16001699: Blood circulation (William Harvey, 1628) 120
Harveys work and influence 120
Evidence for the unified model in 17th century lay practice 120
17001799: Symptom localization (Giambattista Morgagni, 1761) 122
Morgagnis work and influence 122
Evidence for unified model in 18th century lay practice 123
Summary 125
Table of contents

18001899: Tissue cell pathology (Rudolph Virchow, 1858) 125

Virchows work and influence 125
Evidence for unified model lay practice in the 19th century 126
Summary 127
19001990: Medical school standards (Abraham Flexner, 1910) 127
Flexners work and influence 128
Evidence for unified model lay practice in the 20th century 129
Implications for the current studies 130

chapter 6
The main study of two diachronic metaphors of anger 133
Introduction 133
Data collection results 133
The frequency results 134
General trends 135
Implications of the frequency results 136
Comparing the raw frequencies to the scientific advances 136
The discourse analysis 137
The blood and spleen metaphors and prototypicality 137
Analysis of selected historical metaphor samples 138
Chapter summary 160

part iii. Micro-studies of emotion the 19th century

chapter 7
The edge of anger: The spleen metaphor across emotion domains 165
Introduction 165
Purposes 165
Method 166
Data collection 166
Data selection 168
Data analysis 168
Results 169
Discussion 173
Implications of the study 173
Conclusion 174
Metaphor across Time and Conceptual Space

chapter 8
Bubbling happiness: Properties of emotion 177
Introduction 177
Data 177
Research questions 178
Method 179
Materials 179
Data collection 179
Data analysis 180
Results 180
Metaphors of happiness 180
Metaphors of excitation 182
Metaphors of sadness 184
Metaphors of anger 184
Discussion 186
Conclusions and implications of the study 187

part iv. Conclusions and implications

chapter 9
The non-autonomous nature of cognition, language, and culture 191
Introduction 191
Research questions 191
Question 1 191
Question 2 193
Question 3 194
Question 4 194
Implications for conceptual metaphor theory 196
Semantic shift and frequency of use 196
Cultural models as conceptualizations 196
Frequency of use measures and conceptual metaphor 197
Is construal a dynamic process?: Revisiting the issue 198
The embodied core and the cultural periphery 199
The embodied core and the cultural periphery: An example 200
Variation in conceptualization 200
Variation in historical metaphors of emotion: The DME 202
The unity of cognitive domains 202
Conclusion: Conceptualization is static and dynamic 204
Speaker/Hearer interaction and the override 204
Implications for the current study 205
Table of contents

Future research in cognitive-functionalism and conceptual metaphor 205

Dynamic construal 205
Experiential scenes, domain matrices, and non-autonomous
knowledge 206
Cultural models and linguistic forms 206
Diachronic studies of conceptual metaphor 207
The social context of language forms 208
Methodological considerations in cognitive-functional research 208
Multidisciplinary research 208
Corpus size 209
CADS and mixed-methods research 210
Non-linguistic data 210
Additional research techniques 211
Chapter summary 212

Bridging the Gap between theory and real-world language use 215

References 219

Appendix A: Penn-Helsinki corpus 229
Appendix B: ARCHER corpus 231

Index 233
Tables and figures

Table 1 The Historical Four Humors Texts: A Comparison of Basic 100

Table 2 Corpus keyword data collection results 133
Table 3 Metaphor frequency counts, total and by keyword, A.D. 134
Table 4 Spleen metaphor study: Keyword instances, excluded cases, 168
and study cases
Table 5 Bubbling Liquid Metaphors: heat Properties 185
Table 6 Bubbling Liquid Metaphors: pressure Properties 186
Figure 1 Linguistic expressions for anger in American English 43
Figure 2 Elaborations of the fluid CM 45
Figure 3 Elaborations of container destruction in the fluid CM 46
Figure 4 Elaborations of pressure suppression in the fluid CM 46
Figure 5 Elaborations of pressure release in the fluid CM 46
Figure 6 The five stages of the Anger Prototype Scenario 47
(Lakoff & Kvecses, 1987)
Figure 7 Non-prototypical cases of anger: controlled response 50
over time
Figure 8 Non-prototypical cases of anger: intense response over time 51
Figure 9 Two experiential scenes of anger: A comparison 51
Figure 10 Five major scientific advances in human physiology, A.D. 117
part i

Theoretical foundations
chapter 1

The Cognition-Culture interface


This volume describes several research studies in conceptual metaphor (CM) in a

diachronic, longitudinal, and empirical research design across a five-hundred year
time period of the English language. The major purpose of these studies was to in-
vestigate the relationship between the process of cognitive conceptualization (or con-
strual) and non-autonomous knowledge as constituted in cultural models an issue
that has been identified in previous CM research as an important area of study.1 This
chapter will introduce the major theoretical principles that motivate and inform the
research. The first section presents key theoretical constructs that are employed in
the study design and the data analysis. The constructs include cognitive functional-
ism, conceptualization, non-autonomous knowledge, cultural models, usage-based
language theory, linguistic metaphor, and form/meaning pair. An additional con-
struct, time, is discussed in Chapter 3. The second section explicates current issues
in conceptual metaphor research methodology that informed the research design.
The chapter will conclude with an overview of the remaining chapters.

Key theoretical constructs


The research studies presented here combine two traditionally distinct views of
language. One view is that language reflects the way in which the human mind is
organized; this is the cognitive view. The second view holds that language serves a
pragmatic, communicative purpose; this is the functional view. These two perspec-
tives are viewed by some linguists as being in conflict; however, others (including
the studies in this volume) consider them to be complementary (cf. Tomasello,
1998; 2003). That is, cognition and communicative function work together to
comprehend and produce utterances appropriate to the form, structure, content,
and social constraints of a particular language, culture, speech community, and
time period. The implication is that both are essential to understanding language

1. See the section, Cultural Models, later in this chapter for more information.
Metaphor across Time and Conceptual Space

and to the scientific study of language. This perspective on language is termed the
cognitive-functional (hereafter, CF) view. The CF view provides the philosophical
and theoretical principles that inform the research questions, study design, and
analyses of the language data presented in this volume.
The CF theoretical perspective has several fundamental assumptions. The fol-
lowing four principles, generally accepted by the sub-fields research community,
are important for the current study.2
1. The overall goal is to investigate the human mind via language.
2. Language consists of symbolic units, or conventional form/meaning pairings,
which include syntactic, morphological, semantic, pragmatic, and sometimes
phonological information.
3. Meaning is the fundamental characteristic and purpose of language.
4. Meaning is a product of human experience in the world and the patterns of
use of linguistic expressions.
The first assumption shown above, that the purpose of language analysis is to ex-
plore the human mind, is shared among many theorists and researchers in a vari-
ety of research fields, including linguistics.3 In linguistics, the generative grammar
theory developed by Noam Chomsky (1957, 1965) has had a particular focus on
the relation between the human mind and language. The investigation of the mind
(cf. Chomsky, 1975) was one of Chomskys major contributions to the field of lin-
guistics. As a result, since the 1960s cognitive researchers in linguistics focused on
the relationship between language and the mind have contributed important in-
sights to understanding how the mind works and stores information. To accom-
plish this goal, these researchers draw on theory and research in the cognitive sci-
ences, placing linguistics squarely within research activity on cognitive structure.4
Due to Chomskys shift in focus from social context to cognitive structure,
generative grammar historically is not a theory of language function because the
theory focuses on formal grammatical rules, separate from their use by actual
speakers in the real world (Tomasello, 1998). As a result, semantic and pragmatic
aspects of language are often not analyzed in generative grammar research because
semantics and pragmatics are based in the language users idiosyncratic, situation-
al performance, rather than in stable, cognitive competence. In generative gram-
mar, cognitive competence has been the primary object of study.

2. See Geeraerts (2006) for a complete review of these principles.

3. Other fields studying the mind via language include cognitive science, neuroscience, psy-
chology, computer science, artificial intelligence, anthropology and literary studies.
4. Before Chomsky, linguistics was more closely aligned with anthropology and history, fields
which study language in its social context.
Chapter 1. The Cognition-Culture interface

Chomsky further divides the study of syntax into two distinct parts: aspects
that are claimed to be necessary to understand the mind/language relation, termed
the core, and those which are ancillary to that pursuit, termed the periphery. Syn-
tax, morphology, and phonology are core aspects, and semantics and pragmatics
are peripheral. Generative grammar theory partially accepts assumption #2, that
language consists of conventional form/meaning pairs called symbolic units, but
the theory does not accept the related principle that all aspects of language are
important for the study of the mind.5
That semantics and pragmatics, due to their focus on idiosyncratic meaning,
are not accepted for study in generative grammar is made clear in Chomskys fur-
ther rejection of principle #3 meaning is the fundamental characteristic of
language. Generative grammar focuses on the syntactic structure of a linguistic
utterance and what that tells the researcher about the human mind; meaning is
therefore peripheral to that focus. Conversely, CF fully accepts the principle by
collapsing the competence vs. performance and core vs. periphery distinctions to
put all aspects of language on equal footing, and asserts as a fundamental principle
that language is all about meaning (Geeraerts, 2006, p. 3). In contrast to genera-
tive grammar, CF studies all aspects of the form/meaning pair (i.e., phonology,
syntax, morphology, semantics, and pragmatics), in order to understand the hu-
man mind and its capacity for creating meaning.
Not surprisingly, in CF the study of meaning also diverges from the generative
grammar tradition in assumption #4, which concerns how meaning in language is
generated. During the 20th century, theory in semantics focused on formal rules,
including truth conditions, for applying meaning in the real world. Cognitive cat-
egories of meaning were based in knowledge of the objective, physical world in
classical categories. In contrast, cognitive-functional theory employs an alternative
view of cognitive categories of meaning, termed natural categories (Rosch, Mervis,
Gray, Johnson, & Boyes-Braem, 1976). These categories develop through human
experience of the world and the linguistic behavior of the individual language user.
Therefore, meaning develops from an individuals subjective perspective of the
world, rather than from an objective analysis of the organization of the world. In
sum, CF diverges fundamentally from generative grammar theory as well as from
formal semantics concerning the relationship between the mind and language.
The major differences include which aspects of language should be studied, the
role of meaning in language, and how meaning is created. As mentioned previ-
ously, CF accepts all four of the assumptions discussed above, and the current
study accepts them as well.

5. It was on this particular point that generative grammar theory and CF diverged in the late
1960s and early 1970s.
Metaphor across Time and Conceptual Space


As a study in CF, the current research project investigates the general cognitive
process called cognitive conceptualization or construal. Following general CF prac-
tice, the studies described here investigate conceptualization from both cognitive
and functional perspectives. These conceptualizations are the product of everyday
experience in the world: the human mind, situated in a physical body, takes infor-
mation gathered from experience in the world, including the visual, auditory,
tactile, and other biological and perceptual processes and produces conceptualiza-
tions, or meaningful, cognitive construals (i.e., interpretations) of real-world expe-
rience. Linguistic expressions are then employed to express the construals.
The ability which language possesses to express non-linguistic, cognitive con-
ceptualizations has made linguistic data the primary means in CF to investigate
aspects of the human mind within one language or cultural group. Moreover, us-
ing language for this purpose has wider implications: since conceptualization is a
universal cognitive process of human beings, comparing the construals of various
languages has the potential to better understand universal aspects of human expe-
rience, cognition, and language. Though individual human experience is subjec-
tive (grounded in the speaker/writers own personal perspective), some experiences
are seminally important to everyday life, especially experiences of the physiologi-
cal body such as breathing, walking, and expressing emotions. These fundamental
experiences, required for survival in the world, are intersubjective; that is, though
these events are experienced by individuals, the experience is effectively shared by
all human beings and therefore knowledge of the event is essentially the same for
all humans and speech communities. In traditional CF research, conceptualization
serves as evidence for the universal and pre-cultural nature of cognition and con-
strual. Researchers employ language data with the goal to investigate the universal,
intersubjective, cognition-based construals of human experience.6
Conceptualization, as a general process of cognition, produces cognitive con-
cepts that organize human experience in the mind. In CF research, theoretical
constructs for these cognitive concepts that have been proposed include concep-
tual domain and base (Langacker, 1987); idealized cognitive model or ICM (Lakoff,
1987); and frame (Fillmore, 1982).7 The definitions of these constructs overlap

6. See the section later in this chapter, Culture and Cultural Models, for further discussion
of intersubjectivity as it relates to the expression of meaning in a speech community.
7. Conceptual blending (Fauconnier and Turner, 2003) is another type of construal often em-
ployed in CF research, though it is not considered in the current work
Chapter 1. The Cognition-Culture interface

significantly and are often considered isomorphic by CF researchers.8 In this vol-

ume, the term domain is employed when discussing Conceptual Metaphor Theory
(CMT), the field where that term is most prevalent; in discussions of theories or
issues outside of CMT, the term frame is generally used, except where noted.

Non-autonomous knowledge

CF theorists have characterized the long-term knowledge that develops from the
cognitive conceptualization of meaning as encyclopedic and non-autonomous.
Geeraerts (2006) states that [l]inguistic meaning is not separate from other forms
of knowledge of the world [i.e., encyclopedic] that we involves knowledge
of the world that is integrated [i.e., non-autonomous] with our other cognitive
capacities (p. 5; brackets mine). Therefore, different types of knowledge are not
stored in separate, autonomous, cognitive modules but in an encyclopedic network
of cognitive relations, employed to comprehend and to interpret each new experi-
ence in the world. By implication, a person (i.e., an experiencer) brings to bear all
of her previous experience in the world to interpret a new experience.

Culture and cultural models

Some of the non-autonomous, encyclopedic knowledge that an experiencer brings

to a new real-world event is comprised of cultural and social concepts shared by
members of a specific speech community. Geeraerts (2006) notes that languages
may embody the historical and cultural experience of groups of speakers (and in-
dividuals) (p. 5). This type of knowledge is necessary to express construals in a
way that other members of the speech community will understand. The speaker/
writer must tailor the message to fit the non-autonomous knowledge of the hearer/
reader in order to communicate effectively. CF as a research field acknowledges
the role of these types of knowledge including social roles, practices, and the
situational context in conceptualization (Kvecses, 2005; Langacker, 1994).
Cultural knowledge in general is typically defined as local (i.e., relative, non-
universal) knowledge of the social practices in a particular speech community, and
such a definition implicitly assumes that culture has little effect on cross-cultural
universals of cognition and language. However, defining culture as local knowl-
edge obscures the fact that culture also has a universal aspect: each individual has

8. Croft and Cruse (2004, p. 15; see also Croft, 2009) state that base and domain, as defined by
Langacker (1987), are interchangeable terms for each other and also for Fillmores (1982, p. 111)
frame construct. Furthermore, Langacker (1987, p. 150, Footnote 4) equates frame with ICM;
Kvecses (2006, p. 126) and Goldberg (2010, p. 40) concur.
Metaphor across Time and Conceptual Space

an experience of culture in the same way that each person experiences his or her
own physical body. Just as experience in the world (and in turn, cognition) is in-
tersubjective, cultural experience exhibits this characteristic as well (Gibbs, 1999,
p. 153). This idea is not a new one. Keesing (1979) asserted several decades ago
that culture has universal cognitive characteristics: I have recently argued...that
the diversity in thought-worlds in alien cultures has been greatly exaggerated and
that substantial universal cognitive structures very probably underlie cultural
variations (p. 15). If cultural knowledge is comprised of universal cognitive
structures as Keesing suggested, then that knowledge is intersubjective, in the
same way that cognition and language are intersubjective.
All speech communities develop systems of shared cultural knowledge, pro-
ducing perspectives on fundamental conceptualizations that receive detailed spec-
ification in a particular language. These perspectives are organized systematically
in a series of conceptual relations, termed cultural models. Speech communities
and individual members of a community employ these models to interpret em-
bodied experience and determine the meaning of an experience within the com-
munity. Therefore, conceptualization, both in interpreting new experience and in
the linguistic expression of the experience, fundamentally includes non-autono-
mous, encyclopedic cultural knowledge, organized cognitively as systems of cul-
tural models. In this formulation, both linguistic and non-linguistic knowledge
can be employed opportunistically to interpret new experience. Cultural models
exist in all speech communities and are important for effective communication; in
this way, cultural models (as a cognitive construct) are intersubjective shared
across speakers and languages.9
In CF theory, the role of non-autonomous cultural models in conceptualiza-
tion is acknowledged and has been described in detail. For example, in Concep-
tual Metaphor Theory, Kvecses (2005; 2009) has proposed the pressure of coherence
model to account for both the universality and variation of metaphors (2005,
p. 285). In the pressure theory, three systems of non-autonomous knowledge cause
both the universal and variable aspects of linguistic metaphors; these three sys-
tems are (1) experience of the physical human body (also called embodied experi-
ence); (2) social and cultural knowledge (termed context by Kvecses); and
(3) cognitive preferences and styles (2005, p. 285). Social and cultural knowledge
vary in different speech communities and are a product of conceptual metaphor;
it is not a type of knowledge that develops independently (Kvecses, 2009,
pp. 2223). Cognitive styles include a variety of ways in which information is

9. In the current study, the shared nature of a specific, historical cultural model, the Four
Humors model of human health, will be discussed in detail. See Chapter 3.
Chapter 1. The Cognition-Culture interface

processed and manipulated in the human mind.10 Like embodied experience, cog-
nitive preferences are universal (experienced by all human beings) and develop
independently; however, unlike embodiment (and similar to cultural knowledge),
the application of cognitive styles is not universal: Cultures and subcultures may
use them preferentially and to different degrees (p. 286). In the pressure of coher-
ence model, embodied experience is the only type of knowledge that is universal.
In communication situations, Kvecses characterizes the relationships be-
tween these three systems of knowledge as complex and variable each will have
a differential effect on metaphor use in discourse. Universal embodied experience
will have the primary influence on a linguistic metaphor, and cultural context and/
or cognitive preferences can override the universal aspect to various degrees
(Kvecses, 2005, pp. 289292). These differential effects are the result of the com-
municative need (i.e., the pressure) for the linguistic metaphor to be understood
(i.e., to be coherent) within a particular speech community, in light of the specific
content of non-autonomous knowledge in the community.
Kvecses pressure of coherence theory describes in detail the various types of
non-autonomous knowledge that can affect communication, and his theory of dif-
ferential effects characterizes the complex relationships between the three basic
types of knowledge. In addition, the theory views all three types of knowledge as
having simultaneous, complex effects on a linguistic metaphor in discourse. These
are reasonable conclusions based on Kvecses data analysis and his theoretical sys-
tem. However, like traditional views of cultural knowledge discussed previously in
this chapter, in the pressure of coherence model cultural knowledge is viewed as a
type of local knowledge that develops from universal conceptions of embodied ex-
perience. In Kvecses view, cultural models are not independent conceptual sys-
tems they arise out of conceptual metaphor rather than directly from experience.
In contrast, research work by other CF researchers indicates that cultural
models are independent systems of conceptual knowledge that develop directly
out of experience in the world. This view leads to conclusions for conceptualiza-
tion and communication that differ in several significant respects with previous
research on cultural models, including Kvecses pressure of coherence model. This
research is discussed in the next section.

Cultural models as conceptual systems

Cultural models (called cultural units by Goldberg, 2010), as cognitive systems for
organizing knowledge in the mind, have been known and studied in linguistics for

10. Kvecses (2005) identifies a range of cognitive preferences; examples include conceptual
blending, viewpoint preference, and the conventionalization of a metaphor.
Metaphor across Time and Conceptual Space

several decades. As one example, Gentner and Genter (1983) investigated layper-
sons experiential knowledge of the behavior of electricity; the researchers termed
this type of knowledge a mental model, indicating that cultural knowledge is
employed consistently for interpreting lived experience. DAndrade (1987) also
acknowledges the cognitive basis of cultural models in the title of his article, A folk
model of the mind and in the definition provided for the term: a cognitive schema
that is intersubjectively shared by a social group (p. 112; italics mine). More re-
cently, Sharifian, Dirven, Yu, and Niemeier (2008) characterize cultural models as
organized systems of cognitive concepts. These conceptualized systems of cultural
knowledge serve to interpret lived experience (p. 12). In sum, cultural models are
viewed by researchers in various fields as cognitive structures that organize cul-
tural knowledge systematically.
Theorists and researchers in linguistics have also found evidence for the sys-
tematic employment of cultural models or units to interpret, independently, bodi-
ly experience and/or linguistic structure; examples include Perkins (1992), Enfield
(2002), Everett (2005), and Goldberg (2010) for syntax; Shore (1996) for image
schemas; and Cienki (1999), Emanatian (1999), Sinha and Jensen de Lopez (2000),
Niemeier (2008), Svanlund (2007), and Yu (1995; 2009) for conceptual metaphor.
These studies indicate that cultural models have three important characteristics:
(1) a cognitive structure; (2) an organized system of relations; (3) independence
from other cognitive constructs, such as conceptual metaphor.
These characteristics suggest that cultural models have an important role in
both the structure of conceptual metaphor and the creation of linguistic metaphor.
Niemeier (2008) states:
A cultural model, as a representation of a communitys wisdom in a given do-
main, can be seen as both the synthesis and simultaneously as the source of many
clusters of metaphors and metonymies and also of many individual metaphors
and metonymies. In this respect, it differs fundamentally from Lakoff s notion
of conceptual metaphor which sees one global conceptual domain mapped onto
some other domain (p. 350).

The studies discussed above indicate that cultural models are independent, cognitive
conceptualizations, mapping many different cultural concepts within an organized,
multidimensional system. The results further suggest that the conceptual structure
of a cultural model may be more accurately described by Langackers (1987) domain
matrix construct11,12 rather than the conceptual metaphor construct.

11. See Croft & Cruse, 2004, for additional discussion of the domain matrix construct.
12. A parallel notion is Kvecses (2000, pp. 93109) discussion of the state metaphor system,
which he applies to the conceptualization of friendship. Langackers concept was chosen for the
current studies for specific reasons discussed later in this chapter.
Chapter 1. The Cognition-Culture interface

A domain matrix is a complex grouping of cognitive domains (or frames; see

Croft, 2009, p. 14), all of which are related conceptually. The matrix is not consti-
tuted by mappings from many source domains to many target domains; rather, the
matrix is an organized system of domains that are related conceptually to each
other. The relationships are based in both embodied experience and cognitive cul-
tural models.13 Conceptual mappings between domains in two different matrices
will create a conceptual metaphor (or CM); however, a domain matrix is not a CM
(Croft, 2003/1993, p. 178) it is a system of domains related experientially and
conceptually that are employed to create mappings and CM. Domain matrices
(and the individual domains that constitute them) therefore constitute the cogni-
tive basis for both conceptual metaphor and cultural models and the complex rela-
tionships between these two types of conceptualization.
For these reasons, in the current studies a specific domain matrix construct
was employed to account for both the embodied concepts and cultural models that
motivated the linguistic metaphors collected; the specific matrix structure identi-
fied is termed the Domain Matrix of emotion (or DME).14 We argue that the DME
was the dominant conceptualization of emotion in English for three hundred
years, between 1550 and 1850 A.D.

Cultural models as conceptual systems: An example

The following extended example concerning syntax will examine the role of cul-
tural models in interpreting physical experience in the human body for the
purposes of cognitive conceptualization and linguistic expression. Heine (1997)
discusses a key issue facing the cognitive-functional paradigm: the philosophical
conflict between universalism and relativism. Universalism asserts that all lan-
guages can be reduced to simpler underlying principles or cognitive patterns
(p. 10) that are constant across cultures, speakers, and languages, whereas relativ-
ism emphasizes the considerable differences across cultures (p. 11) apparent in
the worlds languages. Heine argues for a middle position, that in addition to a
universalist perspective, there is also need for a relativist perspective (p. 11). He
illustrates this need for a dual perspective in his discussion of the two major types
of deictic orientation that are found in the worlds languages.
Deictic orientation refers to the way in which an experiencer is aligned in
physical space in relation to another object; that is, how the speaker perceives an

13. Langacker states that ...any concept or knowledge can function as a domain..., including
socio-cultural knowledge, such as the conception of a social relationship (1987, p. 63).
14. The DME is discussed in detail in Chapter 9.
Metaphor across Time and Conceptual Space

experiential scene in the physical world.15 16 The two types of deictic orientation
found in the worlds languages are termed face-to-face and single-file, and each
type characterizes a different perspective on viewing a scene and orienting two
separate objects in relation to the speaker. Face-to-face deixis orients the two ob-
jects as facing the speaker i.e., the portion of an object termed the front in the
language is visible from the speakers perspective whereas single-file deixis ori-
ents the objects so that the back portion of an object is visible. Heine provides
examples of linguistic expressions (1997, p. 12) that encode the two deictic orien-
tation types, employing an experiential scene in which a box is placed between the
speaker and a distant hill. (Note: The speaker faces the objects so that both are si-
multaneously within his line of sight).
1. Face-to-face: The box is in front of the hill.
2. Single-file: The box is behind the hill.
In Example 1, deictic orientation characterizes the objects as oriented with the
front-side visible to her, whereas in Example 2 the speaker sees the objects as ori-
ented with the back-side visible to her. The human experience of the scene is the
same across all languages, yet languages which differ in their system of deictic
orientation construe (interpret) the scene in radically different ways. These variant
interpretations are encoded to create two different linguistic expressions with dif-
ferent syntactic constructions and semantic meanings.
Specifically, the face-to-face and single-file orientation types contribute to the
construal of deixis in language because (1) deixis requires the employment of one
of the two orientation types; (2) one type is not privileged over the other in the
worlds languages; that is, either type can be chosen because both are of equal
(i.e., neutral) status from the perspective of embodied experience. These two points
indicate that the specific deictic orientation type employed in a language is selected
by a cultural model because embodied experience alone does not provide the in-
formation necessary to make the choice.17 Without a culturally-specified model to
make the selection, there is no basis for choosing one orientation type over the
other. A cultural model will provide the perspective of the speech community to
select a perspective on the scene that fits the cultural expectations and the linguistic

15. The concept of an experiential scene is a fundamental notion in CF see Chapter 2.

16. The relationship between the experiential scene and the syntax of a linguistic construction
is discussed in detail in Goldberg (1998).
17. The implication that syntactic structures are formed by both cognitive categories and cul-
tural knowledge has been argued by various researchers in syntax. For example, Rauh (2010)
concludes a book-length study of syntactic parts of speech by stating that can be said that
parts of speech represent cognitive categories developed in a particular cultural area during the
course of general education (p. 399; italics mine).
Chapter 1. The Cognition-Culture interface

needs of that community. That culturally-sanctioned perspective then selects the

deictic orientation type which is in line with the cultural models currently in use
in the community, and the choice is encoded in linguistic structure and meaning.
Cultural models fulfill this role to specify aspects of conceptualization that can-
not otherwise be construed solely from physical experience in the world.

Cultural models and syntax: Another example

Other researchers have also found that cultural knowledge influences, even deter-
mines, syntactic structure. Enfield (2002), in a discussion of verb serialization in
Lao, argues that the types of events that verbs typically express (called event typi-
cality) are determined by cultural knowledge. An event must have currency as a
cultural representation shared among members of the speech community, in order
to be employed usefully. Event typicality affects both the production and interpre-
tation of serial verbs, including whether a particular syntactic construction should
be used at all (p. 255), indicating that the use of a construction is determined by
the relevant cultural model. Event typicality therefore is a type of construal which
includes cultural models that specify the use of syntactic constructions in a spe-
cific speech community and communicative situation. In the conceptualized scene,
both the physical characteristics of and the typical interpretation of that scene for
the community are included as specifications.
Enfield (2002), like Niemeier (2008) discussed previously, further states that
cultural knowledge is part of the conceptual system of knowledge:
What we refer to as language and culture can be viewed as part of a mass of
conceptual categories which are shared... it does contain conceptual systems, with
systematic relationships among categories (pp. 232233).18

Enfields analysis can be applied to other aspects of syntax, as well. The two types
of deictic orientation discussed previously are influenced by conceptual unity the
perspective on a conceptualized event that is licensed by the speech community. A
deictic orientation type is selected by a particular speech community depending
on the currency that the type possesses in that community as a culturally-licensed
concept. The interpretation of the scene depends on experiential, linguistic, and
cultural factors. As another example, Goldberg (2010), from the perspective of

18. The view that cultural models are part of conceptual knowledge of the world was argued by
Grace (1987) in his construct of the conceptual event. Pawley (1987) applied this construct to
event encoding for verbs in both Kalam and English. Enfield (2002) used these works as the
basis for his conceptual unity construct to account for verb serialization in Lao. The conceptual
unity of experience in the world and cultural knowledge has been discussed in linguistic studies
of syntax for several decades.
Metaphor across Time and Conceptual Space

Cognitive Construction Grammar Theory, has employed Enfields cultural unity

construct in her own construct of cultural unit (aka, physical experience with cul-
tural knowledge) in her analysis of semantic frames that motivate verb senses in
syntactic constructions. These studies indicate that the construct of conceptual
unity is a useful one for investigating the relationship between cognitive cultural
models and linguistic syntax.

Cultural models provide perspective on a scene

At the end of the section discussing the two deictic orientation types, Heine (1997)
concludes that cognitive patterns i.e., conceptualizations vary not only by dif-
ferences in physical experience, but by differences in culture:
Such findings are remarkable. They give an impression of the wealth of cogni-
tive patterns [italics added] that can be observed in the cultures of the world. No
doubt, such differences must have an impact on the structure of the languages
involved (1997, p. 14).

Heines assertion for a wealth of cognitive patterns indicates that a specific per-
spective on the physical scene (the experiential material for forming cognitive con-
ceptualization) is variable, and variation in perspective is influenced by cultural
models. Cultural models therefore constitute a type of cognitive pattern or con-
ceptualization. In syntax, the cultural model used to construe the scene deter-
mines the deictic orientation type, and then that perspective (i.e., interpretation)
is encoded in language. Via a culturally-sanctioned perspective on the experiential
scene, cultural models imbue the scene with a meaning that can be encoded in a
language and understood by members of that language group and speech com-
munity. As discussed above, such meanings are not available through embodied
experience alone.

The implications of cultural models as cognitive constructs

An important implication of the above discussion is that cultural knowledge is

universal in the sense that it is an inherent component of universal conceptualiza-
tion processes. In addition, cultural knowledge is intersubjective because all
humans experience culture at the level of their own individual perspective
(e.g., viewing a scene), further informed by a speech community perspective via
cultural models. All humans share the same basic process for construing a physical
scene through cultural models; therefore, cultural models are part of the universal
process of conceptualization. In turn, conceptualization affects language structure
and use.
Chapter 1. The Cognition-Culture interface

The above discussion provides evidence from research that cultural models
have the three major characteristics, discussed at the beginning of this section:
(1) cognitive structure; (2) an organized as a system of relations; (3) independence
from other types of construal, including conceptual metaphor. The research pre-
sented in this volume accepts these as fundamental principles.


In sum, conceptualization relies on culturally-sanctioned models for specifying

the particular perspective that will be applied to an experiential scene, and cul-
tural models further contribute to both linguistic structure and semantic meaning
in a specific language. As shown in the example discussed above, cultural models
are a fundamental component of cognitive conceptualization, instantiating a se-
mantic meaning the culturally-licensed relation between an experiencer and ob-
jects in the physical world which is then encoded linguistically in syntax and
vocabulary. These two aspects of conceptualization, embodied experience and cul-
tural models, are mutually interdependent; that is, to describe one independent of
the other is to miss important details of a specific conceptualization. This formula-
tion also follows the CF principle of non-autonomous, encyclopedic knowledge dis-
cussed previously.
Finally, the research conclusions described above indicate that cognitive sys-
tems of cultural models need to be taken into account in studies of conceptualiza-
tion and embodiment, and other researchers have voiced this conclusion regarding
linguistic theory (e.g., Geeraerts, 2010; Geeraerts & Grondelaers, 1995; Gibbs,
1999; Quinn, 1991; Sinha & Jensen de Lpez, 2000). This conclusion follows from
the principle that conceptualization has been found to motivate language form,
including the lexicon, syntax, and metaphor. The major implication of this conclu-
sion is that the universalism vs. relativism debate is a false dichotomy (Domaradzki,
2011, p. 563). CF language theory and research practice should strike a balance
between these two, inherent and independent aspects of the conceptualization
process and construal operations, in order to understand the process and specific
construals expressed in language more thoroughly and accurately. The previous
studies presented in the above section show that this recommendation is currently
being enacted in CF research; the studies presented in this volume are designed to
contribute to this particular research line.

Usage-based theory of language

Taken as a whole, the principles outlined above, including cognitive-functional-

ism, conceptualization, non-autonomous knowledge, and cultural models lead to
Metaphor across Time and Conceptual Space

the conclusion that understanding the structure of language requires investigating

language use in the real world. CF theorists propose a usage-based theory of lan-
guage in which there is an intimate relation between linguistic structures and in-
stances of use of language; specific instances of use are labeled usage events
(Barlow & Kemmer, 2000, p. viii). Cognitive Grammar (Langacker, 1988), and
Cognitive Construction Grammar (Goldberg, 1995; 1998; 2006) are two examples
of major theories in CF that specifically include the usage-based model as a funda-
mental principle.
According to Barlow and Kemmer, usage events over many repetitions lead to
general patterns of use, called schemas, which the members of a speech commu-
nity share (2000, pp. ix-x). Schemas have several important implications for lan-
guage. First, through use, linguistic structures are dynamic and variable rather
than fixed and invariant; language structure is therefore a product of its use A
usage-based model is one in which the speakers linguistic system is fundamen-
tally grounded in usage events: instances of speakers producing and understand-
ing language (p. viii). Second, if usage is crucial to current structure, then usage
both results from and shapes the structure, a kind of feedback loop (p. x).
Usage-based theory prescribes that language will be studied in ways that make the
link between linguistic structure and use explicit rather than implicit, employing a
research method that investigates linguistic phenomena that can be observed em-
pirically. Third, language is investigated most effectively via data of actual usage.
Real language used by real speakers is the only viable method for understanding
language usage, and such data will include its frequency of occurrence: Since fre-
quency of a particular usage pattern is both a result and a shaping force of the
system, frequency has an indispensable role in any explanatory account of lan-
guage (pp. x). Fourth, usage is intimately related to both synchronic patterns of
use and diachronic language change. Fifth, the linguistic system is related to other,
non-linguistic systems, including cultural models (p. xx) and context, both lin-
guistic and social (p. xxi). Finally, there is a relationship between a linguistic
expression and the users comprehension and production of the expression
(pp. xxxii). All of these are important concepts for the current studies and will be
discussed in the remainder of this volume when appropriate.

Usage-based models and empirical research

That usage-based theory is amenable to empirical research is one of the major

advantages of the paradigm. The range of linguistic expressions that can be studied
(Barlow and Kemmer, 2000, ix; cf. Langacker, 1988) include both highly schema-
tized linguistic structures, such as the ditransitive argument structure, as well as
Chapter 1. The Cognition-Culture interface

highly specific forms, such as a particular lexical item or a morphological affix. The
two principles derived from these facts are termed the maximalist and non-reduc-
tionist characteristics of usage-based theory (Langacker, 1988), both of which fit
well with the CF principle to study all aspects of language. Another major charac-
teristic is the bottom-up nature of usage-based theory: the highly-specific, atypical
linguistic structures are privileged over typical, general structures. As Barlow and
Kemmer state, ...the general arises out of the specific, and the specific is what is
most directly taken from experience (p. x). Usage-based theory is therefore ame-
nable to the study of conceptual metaphor which other, more formal theories of
language tend to ignore due to its idiosyncratic nature as well the study of the
effects of cultural models on language, and the study of language within CF theo-
retical principles.19 If general cognitive schemas arise out of specific instances of
experience in the world and language use, then linguistic expressions in actual use
and their attendant cultural models provide concepts needed to investigate cogni-
tive conceptualization processes. For the reasons given in the above discussion,
usage-based models informed the both theory of language and the research meth-
odology for the current studies.

Is conceptualization dynamic?

As mentioned above, Barlow and Kemmer (2000), as a foundational principle of

usage-based theory, state that linguistic expressions have an emergent quality.Stored
cognitive routines recurrent patterns of mental (ultimately neural) activation
(p. xii) are not stored in memory as fixed entities. The stored routines form an
activation network (p. xiii) of neural connections. While the constructs of emer-
gence and activation network are beyond the scope of the current work, the theory
is important for its important implications concerning the production and use of
conceptualizations that form the basis for linguistic expressions. These implica-
tions include (1) the concept of dynamic construal and (2) the role of context in
expressing a conceptualization in a linguistic expression. Each is considered in
turn below.
Theory in cognitive linguistics has recently begun to consider the possibility
of dynamic processes of meaning construction; Kvecses (2005) pressure of co-
herence model is one example discussed previously. Other models include Croft
and Cruses (2004) theory of dynamic construal. They summarize their theory as

19. See Tummers, Heylen, and Geeraerts (2005) concerning specific applications of usage-
based theory to cognitive-functional research.
Metaphor across Time and Conceptual Space

On this view, words do not really have meanings, nor do sentences have mean-
ings: meanings are something that we construe, using the properties of linguistic
elements as partial clues, alongside non-linguistic knowledge, information avail-
able from context, knowledge and conjectures regarding the state of mind of the
hearers and so on (p. 98).

This dynamic nature implies that context, both cultural and situational, has an
important place in conceptualization.
Croft and Cruse (2004) go on to discuss the central place of context in con-
strual. They preface the discussion with a quote from Wittgenstein, who asks why
a linguistic sign seems dead until some unknown factor gives life to it. Croft
and Cruse respond that We shall say that life is breathed into the sign when it is
given a contextualized interpretation...they [signs] are to be distinguished from
the interpretations themselves (pp. 9899; brackets mine). Thus, linguistic ex-
pressions do not possess meaning, rather they are given meaning by a speaker and
hearer in a specific context; the context includes memories, cultural knowledge,
and the current communicative situation. Croft and Cruse liken an interpretation
to a picture or Gestalt: any features are themselves construals (p. 100). Under the
principle of contextualized interpretation, construal is dynamic because the con-
text determines the interpretation of the utterance. The context is not a peripheral
factor in construal, but a central one. For this reason, contextual information is
important for the accurate analysis of linguistic expressions.
Research on the online construal of metaphoric meaning in communicative
contexts lends support to dynamic construal and to the importance of contextual-
ized interpretation. For example, Cameron (2003) recorded discussions in an
elementary school classroom in England on the metaphors used in a science text-
book, studying the comprehension of the scientific topics that the metaphors ref-
erenced. Cameron detailed the difficulties that students had in understanding the
text due to the metaphors used to describe the ozone layer of Earths atmosphere
(e.g., a blanket of gases); in addition, different students created different meanings
for the same metaphor. The problems that the students encountered revolved
around their prior knowledge of Earths atmosphere (the Topic) as well as the rela-
tionship between the Topic and the metaphoric Vehicle (e.g., blanket), which was
often not explicated in the textbook. Cameron concluded that the textual support
that the students needed to use the metaphors appropriately for learning was weak,
leading to misunderstanding and misapplication of the scientific concepts. Cam-
erons study of metaphor in natural language contexts shows that the meaning of a
metaphor is dependent on situational context for accurate construal, and that dif-
ferent dynamic construal operations create different meanings for different indi-
Chapter 1. The Cognition-Culture interface

Other studies have also argued for online, dynamic construal in conceptual
metaphor research. Cameron and Deignan (2006), discussing their construct of
the metaphoreme, argue these metaphoric expressions, with a fixed form and
unique semantics, are created online during verbal interaction. Mller (2008) ex-
plores the gradable nature of metaphoricity, showing that the cognitive activation
of a conceptualization varies for different speakers, hearers, and situational con-
texts. In sum, there is growing evidence that a speaker adjusts a conceptualization
to fit the changing needs of the communicative situation. These studies indicate
that at least some aspects of conceptualization and its expression in a linguistic
form are dynamic and flexible. The principle of dynamism in usage-based theory
has implications for CF research in conceptualization and culture; this issue will be
revisited in Chapter 9, in view of the results of the current research work.

Form-meaning pair

In CF theory, all linguistic expressions combine two types of input: a formal struc-
ture and a semantic meaning, termed a conventional symbolic unit, which com-
bines a linguistic form with a semantic meaning in a form/meaning pair. Tying
form to meaning explicitly follows the CF principle that the purpose of conceptu-
alization or construal is to express a meaning through linguistic structure. The
relationship between form and meaning is tight and inseparable form contrib-
utes to meaning, and meaning contributes to form. CF research practice employs
this principle to design, conduct, and interpret the results of a study, and the cur-
rent study will follow suit.

The semiotic triangle: Form, meaning, and community common ground

Recently, Croft (2008) has proposed a third element in the set of conventional
symbolic units that comprise a language: the speech community which employs a
particular set of symbolic units. Together, the three concepts are combined in a
semiotic triangle of form, meaning, and the speech community in which the
symbolic units are used (Croft, 2008, p. 403). Croft argues that the shared knowl-
edge (termed common ground, following Clark, 1996) employed by a community
of speakers in their use of symbolic units is an important aspect of meaning and
use. That is, the common ground of unmarked cultural knowledge accounts for
the ability of the speaker and hearer to jointly select a salient coordination device
(see next section) which allows for the accurate interpretation of utterances (Clark,
1996), including ambiguous utterances (Green, 1995). Cognitive conceptualiza-
tion is thus a process for categorizing and contextualizing physical experience and
Metaphor across Time and Conceptual Space

cultural knowledge for the purpose of creating a comprehensible interpretation of

meaning within a particular speech community.
Croft demonstrates these principles via the different meanings and uses of the
word subject in the speech communities of linguists, psychologists, British univer-
sity students, and laypersons (p. 403404). The only variable factor is the speech
community in which it is employed, and Croft demonstrates that this variable
changes the meaning and use of the word. The discussion indicates that the com-
munity is an important characteristic of the form/meaning pair in studies of
conceptualization and cultural knowledge. In the current studies of diachronic
metaphor, all three aspects of the form/meaning pair are considered.

Theories of shared cultural knowledge

To understand Crofts point more fully, this section provides details on two theo-
ries of shared cultural knowledge that have implications for the current study:
Clark (1996) on common ground and Green (1995) on the use of shared knowl-
edge to understand ambiguous utterances. First, Herbert H. Clark (1996), a psy-
chologist, discusses the coordination that takes place during communication
between two people. Coordination is a mutual activity that keeps the communica-
tion from breaking down due to misunderstanding; the process involves both par-
ties mutually speaking and listening to utterances but also mutually building
meaning: There must be coordination between what speakers mean and what ad-
dressees take them to mean (p. 325). What is needed to coordinate meaning is a
coordination device that helps determine what the meaning of an utterance is like-
ly to be. Clark uses the example of the Schelling game to illustrate the concept of the
coordination device. In the Schelling game, two people are shown a picture of
three balls a basketball, a squash ball, and a tennis ball. The people (named June
and Ken) are instructed to choose one of the balls, and each is told that a second
person in a different room will also be asked to choose a ball. If both people select
the same ball, both win a prize; if they select different balls, they win nothing.
Clark discusses the outcome of the game and its relation to communication.
June might assume, for example, that she and Ken will both see the basketballs
large size as the clue, focal point, or key that would allow them to coordinate their
expectations and would therefore choose the basketball...if Ken made the same as-
sumption, he would make the same Schelling choice, and they would co-ordinate.
They would have treated this assumed commonality of thought the large size of
the basketball as a co-ordination device (1996, p. 326).

The effective communication of meaning therefore requires the use of a key as the
coordination device that the speaker and the hearer mutually agree is required to
understand the utterance.
Chapter 1. The Cognition-Culture interface

However, there is a problem with coordination devices: how do the speaker

and hearer determine which key is necessary for understanding a particular utter-
ance? There are many potential coordination devices for an utterance, so the pos-
sibility of choosing different keys is high. Clark argues that the principle of joint
salience governs the selection of the appropriate coordination device:
Principle of joint salience: For the participants in a co-ordination problem, the
optimal coordination device is the one that is most salient in the participants
common ground (Clark, 1996, p. 327).

Common ground is the knowledge that both the speaker and hearer share; that is,
the sum of their mutual knowledge, mutual beliefs, and mutual suppositions at
the moment (p. 327). The coordination device for the meaning of an utterance is
chosen from the shared social knowledge of the speaker and hearer; in other
words, joint salience for the key to meaning is determined from the participants
common ground.

Understanding ambiguous utterances

The use of common ground to determine the key to meaning has been employed
by many other researchers, as well. For example, Green (1995) described the pro-
cesses that allow a hearer to determine the meaning of an ambiguous lexical item
or utterance, arguing that ambiguity resolution is important to discourse interpre-
tation (and vice-versa) because the same process governs both. In communica-
tion, polysemy (i.e., different meanings for the same word) causes difficulties for
both ambiguity resolution and discourse interpretation. Green states that rational-
ity is a significant constraint on the resolution of meaning of polysemous words
because What would really be irrational would be using a word to refer to any-
thing except what we estimate our intended audience is likely to take it to refer to,
because it would be self-defeating (1995, p. 11). Similar to Clark, Green proposes
that ambiguity resolution depends on the shared knowledge of the speaker and
hearer; she names this principle normal belief. The principle states that
...the relation normally-believe holds for a speech community and a proposition P
when people believe that it is normal (i.e., unremarkable to be expected) in that
community to believe P and to believe that everyone in that community believes
that it is normal in that community to believe P (p. 11).

Such beliefs are usually unmarked in the utterances of members of the speech
community; for example, as Green points out, though not all members of a
community may believe in a god, members of that society treat one another as
believing in a god except when there is reason to impute the contrary belief to
Metaphor across Time and Conceptual Space

someone... (p. 12). The result is that utterances among members of the commu-
nity assume the normal beliefs of the community, and these beliefs are employed
to interpret the intended meaning of a polysemous word in an utterance and to aid
discourse interpretation in general.

Intersubjectivity revisited

Taking into account these facts concerning shared knowledge in communication,

the definition of intersubjectivity discussed previously needs to be specified in
more detail. The general definition was stated as follows: Fundamental experi-
ences in the physical world are intersubjective; that is, though these events are ex-
perienced by individuals, the experience is effectively shared by all human beings
and therefore knowledge of the event is essentially the same... This definition ap-
plies only to those aspects of human experience that are shared universally by all
humans, such as breathing and walking. However, as previously discussed, cul-
tural models are also universal because all human communities have cultural
models to inform the construal of experience. In addition, the concepts of com-
mon ground and normal beliefs indicate that intersubjectivity applies locally to
the communicative situation between the speaker/writer and the hearer/reader.
Intersubjectivity at the level of a communication event has been known in
linguistics research for several decades. Traugott and Dasher (2002) define it as a
change which results in the development of meanings that explicitly reveal recipi-
ent design: the designing of utterances for an intended the discourse
level (p. 31).20 In this process, called intersubjectification, the speaker/writers
focus shifts from the self (subjective) to the hearer/reader (intersubjective),
prompting changes in the utterance form and meaning that will fit the needs of the
interlocutor. These changes take into account the shared common ground and
normal beliefs of the speaker/writer and the hearer/reader. In sum, intersubjectifi-
cation is employed in speech events at the local level within a particular speech
community to increase the communicative effectiveness of the utterance. Intersub-
jectivity therefore affects communication across languages and speech communi-
ties and also for a communicative event within a particular speech community.


To summarize, the common ground of unmarked cultural knowledge and normal

beliefs account for the ability of the speaker and hearer to jointly select a salient

20. The definition of intersubjectivity employed by Traugott and Dasher has been changing
recently; see Ghesquire and Vandevelde, 2011, pp. 791792, for an overview.
Chapter 1. The Cognition-Culture interface

coordination device which allows for the accurate interpretation of utterances

(Clark, 1996), including ambiguous utterances (Green, 1995). Cognitive conceptu-
alization is thus a process for categorizing and contextualizing physical experience
and cultural knowledge (both of which are intersubjective) for the purpose of creat-
ing a comprehensible construal of meaning within a particular speech community.

Linguistic metaphor

In the current studies, the general definition of this term is taken from Deignan
(2006): A metaphor is a word or expression that is used to talk about an entity or
quality other than that referred to by its core, or more basic meaning. This non-core
use expresses a perceived relationship with the core meaning of the word, and in
many cases between two semantic fields (p. 34). Deignan notes that this definition
is different from the one used in CMT the [linguistic] realization of a cross-
domain conceptual mapping (p. 34; brackets mine). The purpose of Deignans defi-
nition, specifically the core meaning construct,21 is to express the CMT construct in
terms employed in empirical research in linguistics, in order to provide a more pre-
cise, operational definition for linguistic metaphor. This definition forms the basis
for identifying linguistic metaphor in the data collected for the current studies.
The above section described the fundamental theoretical principles that in-
formed the current research studies, except for time, which is presented in detail in
Chapter 3. Below, the implications of the above principles for the current research
studies are discussed.

Methodological issues

Based on the previous discussion of theoretical concepts employed in this study of

conceptualization and cultural models, there are three important methodological
issues that have been discussed in previous research in the area of language and
culture that were also important considerations for the research design of the cur-
rent study. These issues will be presented in order to explicate the reasons for the
methodological choices made for the research described in this volume; the study
design is described in Chapter 4. The three methodological issues are (1) the role
of non-linguistic data in research studies of language and culture; (2) the use of text
corpora for collecting both linguistic and non-linguistic data; and, (3) the applica-
tion of the Corpus-assisted Discourse Studies (CADS) method for analyzing the
data. We will discuss each of these issues in turn.

21. See Allan (2008, p. 22) for a similar construct, termed core concept.
Metaphor across Time and Conceptual Space

The role of non-linguistic data

In recent years, linguists in several sub-fields have specifically considered the

methodological implications for studying the role of culture on linguistic form and
meaning. The context of many of these discussions has often been the study of the
linguistic relativity hypothesis, the proposal that a specific language influences the
way speakers of that language think and view the world. The LRH was first pro-
posed in the 1940s by Benjamin Lee Whorf and has been employed in several dif-
ferent research fields. The theoretical implications of the LRH are beyond the scope
of the current study and will not be discussed further here. However, the method-
ological issues involved in the study of the LRH are instructive for the linguistic
study of language and culture. Two distinct points of view are presented below.
Both views discuss the role of non-linguistic data for delineating cultural models
and other shared knowledge in a given society. After the description of the two op-
posing views, the implications for the current study methodology are discussed.
Lucy (1996) argues that non-linguistic data, such as research on shared cul-
tural knowledge in a speech community, must be collected and analyzed in order
to properly interpret cultural beliefs in linguistic data. He criticizes as inadequate
the reliance on linguistic data alone to investigate the connections between syntax
and culture. He calls this linguacentrism the reliance on the researchers own
linguistic and cultural competence for the accurate analysis of language data. For
Lucy, the solution is to investigate explicitly the connections between language
and culture via the analysis of both linguistic and non-linguistic types of data:
An adequate study of the relation between language and thought should, by con-
trast, provide clear evidence of a correlation of language system with a pattern of
non-linguistic belief and behavior individual or institutional...from a method-
ological point of view, such [linguistic] materials cannot be persuasive by them-
selves...(Lucy, 1996, pp. 44, brackets mine).

Lucy presents a number of studies in which the language-culture connection is

implied but not investigated or is simply ignored. He argues that research methods
which rely solely on linguistic data to understand the relationship between lan-
guage and culture are inadequate to delineate that specific relationship.
Enfield (2000), while concurring with Lucy that non-linguistic data are useful
in research, counterargues that Lucy assumes that language and culture can be
separated effectively in order to study the relationship between them. Enfield
states, is difficult, if not impossible, to isolate anything cognitive or cultural
which is not already imbued with language at a profound level (Enfield, 2000,
p. 126), and language itself is the main data employed for the study of language
and culture, further establishing the inseparability of linguistic form and cultural
Chapter 1. The Cognition-Culture interface

belief. Enfield argues for the use of linguistic data as the primary material for the
study of language and culture, adding that it is unrealistic to demand that studies
concerned with the language-culture-thought relationship should seek exclusively
to demonstrate correlation of a language system with a pattern of non-linguistic
belief and behavior (p. 149); however (slightly revising Lucys terminology),
Enfield agrees with Lucy that the analysis must be done in a way that is non-
linguocentric (p. 150). Overall, the major difference between Lucys and Enfields
positions is the specific role of non-linguistic data in linguistic analysis.

Implications for the current studies

To address this important issue, the current research employed both linguistic and
non-linguistic data, for two reasons. First, accurate interpretation of linguistic data
is especially a problem for diachronic research studies investigating historical cul-
tural knowledge not currently shared or known by the researcher. The analysis of
linguistic data in such studies will therefore be difficult and the conclusions drawn
will be speculative. To deal with this issue, non-linguistic data is used in order to
design a study that is non-linguocentric, the point on which Lucy and Enfield
agree. Second, Enfields position is correct that separating language, thought, and
culture is difficult and this conclusion is consistent with the construct of non-
autonomous knowledge however, the task is not impossible or undesirable; anal-
yses that can do no more than speculate on the nature of a relationship between
variables are of limited usefulness. Therefore, the role of non-linguistic data to
delineate an accurate analysis of historical cultural knowledge is a key method-
ological principle: in order to interpret the language data with increased accuracy
and to provide empirically useful conclusions on the relationship between cogni-
tive conceptualization and cultural models, non-linguistic background data were
collected and used to aid the analysis of the linguistic data. As the discussion of the
results for the current study will show, the non-linguistic data added important
details to the analysis of the linguistic data that would not have been possible by
employing the linguistic data alone, as Lucy points out; in turn, the relationship
between conceptualization and cultural knowledge was more clearly delineated.
Chapter 6 discusses the study results in detail.

Non-linguistic data in conceptual metaphor research

The research tradition in cognitive-functionalism also supports the use of non-

linguistic data. Lakoff and Johnson, in numerous articles and books, have employed
evidence from studies in physiology, psychology, neurology and psycholinguistics
to argue that the results from those research fields provide an existence proof
Metaphor across Time and Conceptual Space

(see Lakoff & Johnson, 1999, p. 38) for the plausibility of Conceptual Metaphor
theory. Extending this position, Fesmire (1994), in explicating the theoretical basis
of cognitive linguistics as a field, discusses the necessity of taking into account the
situational context in metaphor research. He states, A theory of metaphor must
be, in effect, ecological or...pragmatic it must always view human organisms situ-
ated in their social and physical environments (Fesmire, 1994, pp. 152153).
Non-linguistic data is one means of situating humans in their environment, in-
cluding the historical periods studied in the current research.
Finally, the use of such data has previous precedent in the study of conceptual
metaphor. For example, the study of English and Dutch metaphors of anger by
Geeraerts and Grondelaers (1995) employed non-linguistic data, specifically artis-
tic works which employed principles of the Four Humors model, a cultural model
of human health in vogue in English-speaking communities for several hundred
years between the 16th and 19th centuries;22 many other researchers have also
employed non-linguistic data to explicate important theoretical concepts or as
part of the analysis of metaphor data.23 The use of non-text sources provides im-
portant details for the conceptualization that the text samples may not provide on
their own (Eerden, 2009, p. 260).
Conceptual metaphor theory has long acknowledged that there are non-
linguistic realizations of conceptual metaphor (Kvecses, 2010b, p. 63). In addi-
tion, the use of non-linguistic data in conceptual metaphor research is a growing,
though relatively new, practice in CF. Some recent studies have used non-linguistic
data as the basis for the entire research design; this is commonly called multimodal
metaphor research. Examples of conceptual metaphor studies that have employed
multimodal data sources include gesture (e.g., Cienki, 1998), and comics and ani-
mated films (e.g., Eerden, 2009). The multimodal nature of conceptual metaphor,
including non-linguistic data such as artistic works, is a new area of study in CF re-
search, and it has the potential to contribute important details for the conceptualiza-
tion of experience and the cognitive structure of metaphor. In addition, as Geeraerts
and Grondelaers (1995) found, this type of data and its analysis is useful for investi-
gating historical conceptual metaphors and their attendant cultural models.


In sum, CF has long recognized the value of employing corroborating evidence from
other fields to support research results and conclusions. Non-linguistic data can

22. See Chapter 5 for detailed information on the Four Humors model.
23. For example, see Mllers analysis of a political cartoon (2008, p. 5) to illustrate her theo-
retical view of alive and dead metaphors.
Chapter 1. The Cognition-Culture interface

serve as a parallel means of support when describing and accounting for the multiple
perspectives that Fesmires (1994) ecology principle suggests, as well as the study of
language in historical periods that employed different cultural models not known to
the researcher. Finally, the inclusion of this type of data is in line with general CF
research practice and has been employed in a previous study of conceptual meta-
phor. For these reasons, the analysis of non-linguistic data was included in the re-
search design of the current study of historical metaphors of anger in English.
Specifically, in the current study non-linguistic data are important for under-
standing the role of culture in a particular data sample, as well as the influence that
culture has on the semantic meaning of a metaphoric expression and/or the CM that
motivates that expression. Without non-linguistic data, understanding historical
cultural models and delineating their influence on conceptualization leads to specu-
lations, rather than empirical conclusions, concerning the relationship between
cognition and cultural knowledge. Understanding these issues, non-linguistic data
were included in the research methodology for the current study.

The use of text corpora

Introspection as a data collection strategy

The second important methodological issue involves text corpora, particularly

texts collected in digital (electronic) media. In CF research, the typical analysis
technique for collecting language data is introspection. The researcher, using her
own competence in the language under study, creates one-sentence data samples
that employ the form and meaning features that are the target of the study. The
method is a valid one in synchronic research to demonstrate the structural param-
eters of a particular language form under study, and for this reason, introspection
has been the basic data analysis technique for the study of contemporary language
structure in CF research.
However, due to the complex issues involved in creating useful and valid ex-
amples of specific features, the data samples are almost always single sentences,
which preclude the study of extended discourse beyond the sentence level; more-
over, cultural models are difficult to discern because the ecology the social, cul-
tural, historical, and situational context of the form is not included. Without the
contextual information, in many cases the effects of cultural models on language
use and change cannot be discerned with any accuracy.

Introspection and the problem of context

The general strategy to collect and analyze single sentences also creates another
problem specific to research in conceptual metaphor. Croft (2003/1993) argues that
Metaphor across Time and Conceptual Space

accurate identification and analysis of a CM, including the conceptual domains

involved, are often not decidable from the elements of the sentences themselves.
The domain in which a predication is interpreted can be determined by context...
which need not be overtly expressed in an utterance (p. 199). Furthermore, with-
out contextual information, interpretation in any domain is possible, short of
semantic incompatibility (and conventional limitations on the figurative interpre-
tations of particular words and phrases) (p. 199). Introspective analysis may lead
to inaccurate conclusions about the conceptualization that motivates a linguistic
expression, due to the lack of contextual information related to the specific concep-
tual domain(s) and the cultural model(s) involved. The data samples produced by
introspection reflect the unique, individual knowledge and usage of the researcher
rather than the shared, collective, contextualized knowledge of the target speech
community. An important implication is that generalizing the results of introspec-
tive analysis may lead to inaccurate and even inappropriate generalizations across
speakers within a community and cross-culturally.
These weaknesses have usually limited research in Cognitive Linguistics to
word-level and sentence-level analyses and their implications for cognitive con-
ceptualization, through an appeal to universal bodily experience via the universal
aspects of intersubjectivity. Though this approach is defensible logically, in practi-
cal terms the data analysis of the samples often reduces to inference the contribu-
tions that situational context and cultural models bring to the conceptualization.
Geeraerts and Gevaert (2008, p. 339) assert that employing preselected examples
to fit a particular theoretical model leads to, at best, speculative conclusions.
Sinclair (1991) concurs:
...the contrasts exposed between the impressions of language detail noted by
people, and the evidence compiled objectively from texts is huge and systematic.
It leads one to suppose that human intuition about language is highly specific, and
not at all a good guide to what actually happens when the same people actually
use the language (p. 4).24

In addition, usage-based theory asserts the methodological importance of situ-

ational context to understand language use by individual speakers and by

Text corpora as a data collection strategy

There have been discussions in the CF field concerning the limitations of intro-
spection for data analysis and interpretation (e.g., Croft, 1998; Gibbs, 2006), as

24. Partington (2004) is even more pointed, quoting Leonardo da Vinci There is nothing
more deceptive than to rely on your own opinions... (p. 13).
Chapter 1. The Cognition-Culture interface

well as the practical limits of current CF methodology for the study of the human
mind (see Casasanto, 2009, Peeters, 2001, and Sandra, 1999, for critiques). Some
researchers have called for employing text corpora to address the weaknesses in
the introspection method (Geeraerts & Gevaert, 2008), and others argue for com-
bining corpora with introspection in one study (Kvecses, 2011) to increase the
power of the analysis. Despite the increasing use of corpora in CF research, intro-
spection has its place as a research method, but data beyond the researchers own
linguistic competence are needed to corroborate the findings and to delineate im-
portant linguistic details.
In addition, the usage-based model of language, increasingly prevalent in CF,
fits well with corpus study (Grondelaers, Geeraerts, & Speelman, 2006, pp. 149150).
Corpora provide data of authentic, natural language use from a wide variety of
language users, allowing researchers to study the actual (not presupposed) uses of
linguistic forms. In recent years, corpus study in combination with usage-based
models has been employed with greater frequency in CF research.

Corpus research in Cognitive-Functionalism

As mentioned above, new research methodologies are being developed that spe-
cifically address the limitations of the introspection method. Some CF researchers
employ compiled corpora of language data in their investigations (e.g., Deignan,
2005; Grondelaers, Geeraerts, & Speelman, 2006; Heylen, Tummers, & Geeraerts,
2008; Stefanowitsch & Gries, 2006; Oster, 2010). A compiled corpus employs the
general principles of scientific research to randomly select texts in the target lan-
guage to create a collection of natural language texts that are representative, at a
scientifically valid level, of the form, meaning, and use of language of the target
speech community. Through the systematic compilation of a corpus, the data sam-
ples collected reflect a wide variety of text modes (e.g., written, spoken), registers
(e.g., formal, informal), and genres (e.g., newspapers, research articles, speeches).
As Biber, Finegan, and Atkinson (1994) note, in compiled corpora [t]he general
goal has thus been to represent as wide a range of variation as possible (p. 4).
Biber, Conrad and Reppen (1998) further demonstrated that analyzing texts from
only one or two registers or genres leads to inaccurate generalizations concerning
the use of words and grammatical structures in language. They state, ...a corpus
restricted to any one register will not represent language use in other registers
(p. 34). In contrast, corpora compiled on scientific principles, which are specifi-
cally designed to be representative of language use across registers and genres, af-
ford a scientifically valid view of the language under study as it was actually used
in a historical period across texts, genres, native speakers, registers, and com-
municative functions.
Metaphor across Time and Conceptual Space

In addition, the structure and use of a linguistic expression can also be quanti-
fied because the total number of words in a compiled corpus can be calculated and
various statistical measures can be applied, offering empirical data for analysis. In
the current study, two compiled corpora were employed to view changes in meta-
phor form, meaning, and frequency of use over time. The results of the analysis of
language use are more easily generalized to the behavior of native speakers, due to
the use of corpora which are highly representative of the historical language form-
meaning pairs under study.
Finally, the particular corpora selected can provide contextual data to investi-
gate the contributions of cultural knowledge to conceptualization. Corpus research
methodology has been criticized for focusing on short passages (often word strings
of 80 characters or less) that limit the amount of situational context that can be em-
ployed in the linguistic analysis (Hunston, 2002), but more recent studies include
longer word strings. Corpus study is amenable to large samples that include suffi-
cient contextual data for the analysis of cultural models. In addition, recent studies
in CF have discussed the usefulness of corpora to analyze extended discourse to
delineate synchronic semantic and pragmatic features (e.g., Oster, 2010; Sim,
2011). The current study addresses these issues by selecting corpora and corpus
analysis systems that include complete contextual data the historical time period,
the author, the title of the work, the title of the publication, the publisher, and the
words, sentences, and paragraph(s) in which a linguistic expression is placed.
The collection of data samples which includes extended context allows for a
more accurate analysis of the meaning of each data sample as well more accurate
identification of the cultural model(s) shared by the speaker/writer and the in-
tended audience. Based on these considerations, corpus research methods were
viewed as having the potential for delineating the effects of cultural models on
linguistic metaphor data over time.

Corpus methods in perspective

Of course, all research methods have specific weaknesses. There are several issues
in corpus research that relate to CF research: (1) the CF theory of language, (2) the
use of frequency statistics, (3) corpus text selection, and (4) short-term language
forms vs. long-term language patterns. Each of these will be discussed below.
1. The CF Theory of Language. First, from the general view of research philosophy,
corpus study on its surface seems to be a poor fit for the non-objectivist, inter-
subjective theory of language employed in the CF field. However, Grondelaers,
Geeraerts, and Speelman (2006) argue that empirical data, and corpus data in
particular, are not antithetical to CF theories or research principles:
Chapter 1. The Cognition-Culture interface

[i]f we assume that language is not genetically so constrained as to be uniform

all over the globe, and that linguistic communities are not homogeneous (two as-
sumptions that would seem to be congenial to the non-objectivist stance of Cog-
nitive Linguistics), then a broader empirical basis than ones own language use is
necessary to study the variation...the cognitive phenomena that you study may be
non-objectivist, but you do try to study them objectively (2006, p. 150).

The researchers also acknowledge that data must be interpreted, as in any non-
objectivist research design, but the results from corpus studies will lead to
more accurate and refined hypotheses than data generated from introspection
alone. In sum, objective corpus data can be studied and contribute to deeper
understanding of the non-objectivist theory of language employed in CF; the
current study accepts this important principle.
2. The Use of Frequency Statistics. Some CF researchers have cautioned against
overemphasizing the effects of the frequency of occurrence of a linguistic form
as evidence to support the overall significance of that form (Gries, Hampe, &
Schnefeld, 2005, p. 665).25 In conceptual metaphor research, a parallel con-
cern is the general view that frequency data do not aid the analysis of specific
aspects of a conceptual metaphor, in particular its prototypicality (Kvecses,
2008, p. 200). The theoretical reason for this view is that a conceptual meta-
phor is a construct of the human mind, not a linguistic form (Lakoff, 1986;
1993); therefore, the number of linguistic forms that may potentially express a
single conceptualization is highly variable, reducing the usefulness of frequency
measures for describing a conceptual metaphors prototypicality. The current
studies also support this theoretical view. In addition, we agree that overstat-
ing the significance of any data analysis technique is a valid concern in re-
search practice. However, we argue that careful research design and analysis
should reduce significantly the potential negative effects of any technique, in-
cluding frequency of use measures.
Frequency of use data provide information useful for conceptual meta-
phor research, such as the identification of lexical items that indicate the
cognitive mappings between the source and target domains, and also identifi-
cation of the cultural models that provide perspective on the experiential
scene. For example, the lexical item blood as a cognitive source domain has
been found in many studies to map to the target domain of ANGER. While
this result was first found in synchronic introspection studies, in diachronic
research this specific mapping can only be confirmed by frequency analysis in
historical corpora. Introspective data for a historical form would be highly
inaccurate, as the previous discussion on non-linguistic data showed (the

25. But see Bybee, 2010, p. 98, for a detailed defense of mere frequency as an analysis tool.
Metaphor across Time and Conceptual Space

effects of linguacentrism). Diachronic studies of metaphor must therefore rely

on the historical linguistic form, including lexical items, to investigate cogni-
tive conceptualization.26
In addition, as discussed previously, usage-based models of language re-
quire frequency data to determine the actual use of a linguistic form for com-
munication within a speech community, an important consideration in studies
of conceptualization and cultural models; without frequency data, the use of the
form in a speech community is more difficult to characterize accurately. In light
of these issues, in the current study, the frequency of a linguistic expression was
a factor in the quantitative analysis of the historical conceptual metaphors un-
der study as well as the analysis of cultural models. Conclusions are then drawn
concerning the specific conceptual metaphors under investigation.27
3. Corpus Text Selection. The proper selection of texts is an important method-
ological issue in corpus research. Trim (2011), regarding the role of culture in
diachronic studies of language, rightly points out that frequency can be a local
or a short-term phenomenon:
One particular frequency count may only reflect one section of the language
community. In addition, single historical events may cause salience in particular
items to increase considerably over a short period of time. It is this issue that ren-
ders frequency counts difficult in the analysis of general long-term trends in the
evolution of diachronic salience (Trim, 2011, p. xii).

As discussed previously, the general principle that underlies corpus selection

is the principle of representativeness; that is, the degree to which the corpus
reflects the patterns of use of the target speech community. To increase the
representative nature of the corpus, the texts must be selected in a principled
way that incorporates the scientific method, such as selecting texts via random
sampling. The samples are taken from a wide variety of text genres (e.g., news-
papers, magazines, personal letters and diaries, and others), authors, geo-
graphic regions of the target speech community, and time periods. By this

26. This same methodological principle can also be applied to synchronic studies of metaphor,
though for a different reason. While in synchronic research the researcher can usefully employ
introspection to identify the source domain and the mapping to the target domain, generalizing
the results across the members of a speech community (and cross-linguistically) requires confir-
mation from language data that are independent of the researchers own non-autonomous knowl-
edge. Following this procedure will also limit the effects of researcher bias. Frequency of use
statistics from data collected in corpora can be employed to confirm the results of the introspec-
tion analysis. For the reasons discussed in this section, we recommend the use of corpus research
designs and frequency of use statistics for both diachronic and synchronic studies of metaphor.
27. The use of frequency data for investigating conceptual metaphor is discussed in more detail
in Chapter 4.
Chapter 1. The Cognition-Culture interface

careful procedure, the corpora used in a research study will be more highly
representative of the speech communitys linguistic structure and actual usage
patterns. Trims valid concern about text selection is met by the current, stan-
dard research methods employed in corpus linguistics.
4. Short-term Language Forms vs. Long-term Language Patterns. This issue has
also been discussed by CF researchers. For example, Kvecses (2011) states
that corpus studies tend to find irregularities in language use rather than
regular patterns. It is true that corpora delineate irregularities, but this is not a
limitation of this type of research. Corpus study can bring out regularities,
especially in diachronic research which gather data samples over many years;
regular patterns emerge across a large time span. Discovering regular patterns
in corpus data is not qualitatively different from finding patterns in introspec-
tive data it depends chiefly on the design of the corpus and the procedures
employed for data analysis. The goal to uncover regular patterns in corpora is
partly met by careful text selection. A corpus that is designed to be highly
representative of a specific speech community will effectively limit, if not elim-
inate, the occurrence of short-term effects and linguistic fashions in the
whole corpus. This is especially true in diachronic studies because long-term
historical time periods increase the chance that long-term linguistic patterns
will be found; a study of several centuries increases the trend toward finding
such patterns, compared to a study of a few years or decades. A diachronic
study design over five centuries of data in corpora, as employed in the current
study, is therefore an advantage for collecting data that reflects long-term,
regular linguistic patterns.
The other factor that limits the effects of short-term historical trends on long-
term patterns is effective procedures for data analysis. As discussed previously,
in a non-objectivist approach, objective data must still be interpreted in sys-
tematic and scientifically-valid ways (Grondelaers, Geeraerts, & Speelman,
2006, p. 150). In the current study, the combination of non-linguistic data of
the historical period to interpret the samples, coupled with systematic collec-
tion and detailed analysis of the linguistic data, reduces the chances that the
results of the data analysis will be unduly influenced by short-term linguistic


Overall, current corpus research methodology provides both practices and tech-
niques that can achieve scientifically-valid results. Though Trims (2011) and
Kvecses (2011) concerns are important considerations, current standard practice
in corpus linguistics has addressed these specific issues in detail and has developed
Metaphor across Time and Conceptual Space

methods to limit and even eliminate the effects of confounding influences in text
corpus research studies. Like the introspection method, corpus study has its own
advantages, and these advantages have the potential to contribute significantly to
our understanding of language; the weaknesses of corpus research can be dealt
with effectively with good research design and careful analysis.

Corpus-assisted discourse studies (CADS)

We have argued above that the extended discourse context in which a linguistic
metaphor is placed is needed in order to develop accurate identification of the
conceptual domains and the historical cultural models (unknown to the researcher)
that determine important aspects of the conceptualization. Therefore, collecting
extended discourse data in corpora for analysis was an important methodological
requirement of the current work. In addition, statistical measures were needed to
study conceptualization and cultural models empirically. However, according to
Partington (2004), discourse analysts have not generally used corpora in their
studies of discourse, and corpus researchers have not generally used discourse
analysis methods in their studies of corpora (pp. 1112). These two research tradi-
tions have rarely been used in concert, but a research method that combines both
corpora and discourse methods has potential to provide the advantages of both.
The use of compiled corpora to conduct both quantitative and qualitative
analyses sometimes called a mixed research method is, at a philosophical
level, the ideal in corpus studies (Biber, et al., 1998). Recently, a new research par-
adigm, Corpus-Assisted Discourse Studies (CADS), has been proposed to fill the
methodological gap (see Partington, Morley, & Haarman, 2004). The studies in
Partington, et al. were designed to fit between the quantitative and qualitative
methodological poles, attempting to combine the advantages of each to create
more powerful analyses of linguistic data.
For CF in particular, this approach addresses the theoretical principle that lan-
guage is comprised of form-meaning pairs, and also addresses Crofts (2008) pro-
posal that research must take into account the shared knowledge of the speech
community available in contextual data. Several other CF researchers have also
recommended combining quantitative corpus study with qualitative introspective
or discourse research designs for the purposes of studying conceptual metaphor,
including Geeraerts, Gevaert, and Speelman (2011), Kvecses (2011), and Sim
(2011). The CADS approach provides a principled method for combining corpus
and discourse research methods; for this reason, the CADS method was employed
in the current study. The following section provides a description of the CADS
Chapter 1. The Cognition-Culture interface

The CADS Method: An Example. To briefly summarize the CADS method, the
following steps are outlined by Partington (2006): (1) develop a research question;
(2) choose or create a corpus for data collection of linguistic items; (3) choose or
compile a reference corpus (for comparison to the main corpus created in Step #2);
(4) compile frequency lists of keywords and conduct comparisons of the key words;
(5) determine the existence of multiple instances of key items; (6) investigate the
context of use of the key items in the corpora. Partington (2006) describes a series
of studies which demonstrates the CADS methodology and its usefulness for re-
search in cognitive linguistics. One of the studies investigated the use of concep-
tual metaphor in political discourse during the Clinton presidency (Step #1 in the
method outlined above). Steps #2 to #6 are described briefly below for that study
as an example of the CADS research method.
After compiling a digital corpus of 250,000 words of Clinton-era White House
press briefing transcripts, frequency lists of words in the corpus were created. The
frequency lists revealed a pattern of prepositions or adverbials used to describe
forward movement, such as forward, forwards, to, and towards (Partington, 2006,
p. 275). Partington hypothesizes that these orientational metaphors28 have a con-
ceptual relationship with the verb move because the keywords often appear with
the verb in various clusters of four contiguous words; examples include as we move
forward; to move forward with; continue to move forward (p. 276). This systematic
association between the keywords and the verb was hypothesized to be motivated
by two conceptual metaphors, progress is forward motion and moving
forward is necessary.29
Discourse analysis of the situational context in which the clusters appear in the
corpus revealed that the conceptual metaphors include other important concep-
tual dimensions, all of which relate to the role of the presidency in U.S. politics:
The administration must at all times be seen to be making progress, to be moving
or headed in the right direction (p. 276); that is, positive progress is required for
the president to fulfill his political role. Negative progress, found in the corpus in
the keywords such as backwards, backpedaling, and bogged down, is to be avoided.
CADS and Conceptual Metaphor Research. It is clear from the description of
the study results that the combination of the quantitative frequency lists and qual-
itative discourse analyses reveals the nature of the conceptual metaphors in greater
detail than either analysis type would have accomplished alone. The results of

28. As Partington notes, Lakoff and Johnson (1980) state that prepositions and adverbials are
types of orientational metaphors that reflect the conceptualization of the movement of the phys-
ical body in space.
29. This volume uses small capitals, the standard CF notation system, to denote a cognitive
conceptualization, generally in the form abstract concept is concrete entity.
Metaphor across Time and Conceptual Space

Partingtons metaphor study indicate that the CADS approach provides a method-
ological solution for the study of form-meaning pairs of contextually-situated dis-
course, including metaphor. Two other aspects of the current study, its diachronic
design and the requirement to include non-linguistic data, also must be consid-
ered before adopting the CADS method.
CADS and Diachronic Study Designs. Concerning diachronic study, Parting-
ton (2011) has recently served as editor for a special issue of the journal Corpora
which specifically applied CADS to diachronic study.30 The method was found to
be amenable to fit the design of the current diachronic research studies.
CADS and Non-linguistic Data. As for the analysis of non-linguistic data, Par-
tingtons metaphor study of political speech, discussed above, considers only con-
temporary language use that he, as a native speaker of English, can analyze using
the knowledge of current cultural models that he shares with other members of
the speech community; non-linguistic data is not specifically included in the
CADS method. Conversely, data types in CADS are flexible in that both quantita-
tive and qualitative data are specifically included. Non-linguistic data is a type of
qualitative data, so including it in CADS is appropriate.
In sum, both issues diachronic study designs and non-linguistic data can
be incorporated into the CADS study design without major changes to the funda-
mental principles or the research procedures. For the current diachronic studies of
historical metaphor, CADS was employed in the research design, including the
data collection and analysis procedures.

Chapter summary
The methodological issues concerning the use of non-linguistic data, text corpora,
and the CADS method are complex and must be handled in a principled way to be
used effectively in research design. However, all three techniques were found to be
useful for the goals of the current diachronic studies of conceptualization and cul-
tural models.

Plan of the volume

The plan of the remaining chapters in this volume is as follows. Chapter 2 provides
an analysis of Lakoff and Kvecses 1987 study of the Conceptual Metaphor (CM)
of anger. The results of the analysis leads to a proposal for two distinct and unique
conceptual metaphors: the blood metaphor and the spleen metaphor. The chapter
also discusses the previous research literature that pertains to the synchronic study

30. The studies covered a brief time-course, however, from 1993 to 2005.
Chapter 1. The Cognition-Culture interface

of conceptualization and cultural models. Chapter 3 discusses diachronic research

and its advantages for the study of conceptual metaphor. Previous historical re-
search in conceptual metaphor is discussed in detail, which provides the underly-
ing principles that guided the design of the current studies in diachronic metaphor.
Chapter 4 details the specific research questions and the research methodology
employed in the current studies. Chapter 5 details the results of the analysis of the
non-linguistic data of historical cultural models; this information provides the ba-
sis for the analysis of the linguistic samples of historical metaphors of anger.
Chapter 6 provides the results of the macro-study of the blood metaphor and the
spleen metaphor over five centuries, including the meaning and use of the two
proposed prototypes, using the non-linguistic data to aid in the interpretation of
the linguistic data. Chapters 7 and 8 describe two additional, micro-studies of
diachronic conceptual metaphor in the 19th century on questions related to the
results of the five-century study. These chapters provide further details on the
spleen metaphor and related conceptualizations of emotion. Chapter 9 summa-
rizes the results of the three studies and proposes the Domain Matrix of emotion
as the major locus of emotion concepts and attendant cultural models. In addition,
the implications for conceptual metaphor theory, research methodology in cogni-
tive-functionalist linguistics, and future research are discussed. Finally, the
Epilogue presents a brief justification for applying the results of the current studies
to the teaching of metaphor in the second language classroom.
chapter 2

Diachronic aspects of synchronic concepts


The purpose of this chapter is to provide the analytical foundation for the study of
the effects of diachronic cultural models on synchronic conceptualizations of emo-
tion. In Chapter 1, it was noted that CF researchers acknowledge the relationship
between diachronic forms and their synchronic counterparts. This chapter extends
that idea to investigate the influence of historical cultural models on the process of
cognitive conceptualization in synchronic language use.
The method for this investigation is a detailed analysis of the prototypical CM
of anger, as presented in Lakoff and Kvecses (1987), which follows an introduc-
tion to Conceptual Metaphor Theory (CMT). The analysis of the 1987 study
concludes that certain aspects of the CM of anger indicate the influence of dia-
chronic cultural models, specifically the Four Humors model of human health;
this conclusion supports the research studies discussed in Chapter 1 (see the sec-
tion titled, Cultural Models).
In addition, an analysis of the non-prototypical cases of anger identified in
Lakoff and Kvecses (1987) reveals a second anger prototype scenario; this sce-
nario is markedly different from the one described in the 1987 study, leading to the
preliminary conclusion that there is a second prototypical CM of anger. To differ-
entiate the two, Lakoff and Kvecses prototype is termed the CM of blood, and the
second prototype is termed the CM of spleen (hereafter in this volume, the blood
metaphor and the spleen metaphor, respectively). The implications of these results
are discussed concerning conceptual metaphor theory in general and the current
research in particular. Research studies on synchronic conceptual metaphor are
presented which support the proposed relationship between conceptualization and
cultural models. (Diachronic studies of conceptual metaphor and cultural models
are discussed in Chapter 3.) Finally, a summary of the chapter is presented.

Introduction to conceptual metaphor theory

Conceptual metaphor (CM) is a construct developed by George Lakoff and Mark

Johnson (1980, 1999). A CM is not a linguistic form; as Lakoff (1986) explains, it
Metaphor across Time and Conceptual Space

is a figure of thought a cognitive structure rather than a linguistic one. CMs

are formed by a persons experience in the physical world (termed embodied real-
ism, or bodily experience; see Johnson, 1987). Lakoff and his colleagues assert that
conceptualizations provide the cognitive structure for interpreting new experi-
ence. For example, when a person expresses anger, his or her body becomes warm,
the skin turns red, and the person shakes the fists at the person or circumstance
that has caused the anger. Over many repetitions of the experience, a CM called
anger is a hot fluid in a container1 (Lakoff, 1987; Lakoff & Kvecses, 1987)
forms and becomes part of the cognitive structure in the mind of the experiencer.
Essentially, the CM maps the abstract concept of anger (the target domain) onto
the concrete entity of a container of hot fluid (the source domain), which refer-
ences the human body and its fluids, such as blood).2 The hot fluid CM is in turn
used to interpret new experiential situations; for example, when a friend becomes
angry during a conversation, a metaphoric expression, such as His blood boiled,
is employed to interpret the experience via the CM. In this way, cognitive concepts
are formed by repeated everyday experience and are later expressed through lan-
guage to interpret new experience.
For Lakoff, et al., embodied experience is the basis for conceptualization, and
conceptualization in turn creates conceptual domains, and a mapping between a
source and a target domain creates a CM, and the CM motivates both the form and
semantic meaning of a linguistic expression. For cognitive-functional researchers,
this process is characterized by the principle, conceptualization equals semantic
meaning (Langacker, 1987). In addition, some conceptual metaphors, such as The
hot fluid CM, are considered by Lakoff, et al. to be ubiquitous across languages
and cultures, serving as evidence for the universal and pre-cultural nature of cogni-
tive conceptualization, meaning, and CMs.
The current study accepts that CMs have cross-linguistic aspects, but applies
the principle equally to cultural knowledge at the level of cognitive conceptualiza-
tion of embodied experience by experiential scenes (see Chapter 1). We argue that
cultural knowledge is employed simultaneously with embodied experience in the
development of cognitive domains and the mapping of domains to produce con-
ceptual metaphors in the mind. We discuss the reasons for this view below.

1. This volume employs small capitals, the standard CF notation system, to denote a cognitive
conceptualization, generally in the form abstract concept is concrete entity.
2. As discussed in Chapter 1, the basic types of construal identified in cognitive-functional
research, including domain, base, idealized cognitive model, and frame, are the cognitive basis
for the development of CM. A CM maps pre-existing non-autonomous knowledge of experi-
ence, including cultural models, constituted in cognitive conceptual domains.
Chapter 2. Diachronic aspects of synchronic concepts

Perspective on a scene via cultural models

As discussed in Chapter 1 regarding the example of deictic orientation, a specific

perspective of a experiential scene perceived by an experiencer situated in a par-
ticular situation and speech community indicates that the conceptualization in a
linguistic expression requires a specific perspective on the scene (e.g., single-file
deictic orientation) that instantiates a particular syntactic structure and a seman-
tic meaning, creating a form-meaning pair that is not neutral with regard to non-
autonomous cultural models. Cultural models are therefore required to make the
choice of a specific intersubjective perspective for the same reason that CF theory
privileges meaning over syntax: communication of meaning is the primary goal of
language. In order to communicate effectively in a particular speech community,
the form/meaning pair selected by the speaker will include cultural knowledge
shared by the speaker and the hearer.
If this analysis is correct, cultural models operate at the level of cognitive con-
ceptualization. The universal aspects of cognition include embodied experience in
the real world, cognitive conceptualization and construal, and cultural models.
Recall from Chapter 1 that Croft (2008) has argued that linguistic form, semantic
meaning, and community knowledge are all required for communication between
a speaker and a hearer in a particular speech community. The specific details of
these three aspects of human embodied experience are influenced by the local
speech situation, local bodily experience (the individuals subjective experience),
and local iterations of cultural models, which in turn influence the specific details
of the conceptualization and its employment in linguistic expressions. An experi-
ential scene incorporates both an experiential component and a cultural compo-
nent. The experiential component perceives and conceptualizes the scene in the
physical world, but the basis for selecting an intersubjective perspective on the
scene (e.g., whether an object faces toward or away from the perceiver; whether
heat is the primary dimension of anger) is supplied by the cultural model.
A neutral, pre-cultural perspective on the scene, often assumed in conceptual
metaphor theory, is not sufficient to express a conceptualization in language; the
meaning and significance of the scene for the speaker and hearer must also be
encoded, and cultural models fulfill this function. Therefore, both embodied expe-
rience and cultural models are necessary for cognitive conceptualization (construal
of the scene) and for forming CM.3

3. This hypothesis in the current study is consistent with the CF principle of non-autonomous
knowledge, discussed in Chapter 1, in which experience in the world and social/cultural iden-
tity are necessary and inseparable aspects of experiential knowledge and meaning expression.
Metaphor across Time and Conceptual Space

Analysis of the CM of anger

An analysis of Lakoff and Kvecses (1987) seminal study of the CM of anger is

provided in the next section to demonstrate the issues discussed above. This study
is often cited to support the conclusion that the conceptual metaphor described
in the 1987 study is the prototypical conceptualization of the human experience
of anger in English and a wide variety of other languages (cf. Kvecses, 2005;
2008). A review of the study and its chief results are followed by an analysis of
selected data. The analysis identifies two separate CMs of anger via differences
in the conceptualizations that motivate specific metaphoric expressions of anger.
This result indicates that each conceptualization is motivated by a separate expe-
riential scene, conceptual domains, and mappings of those domains, creating two
separate CM.4

The bodily experience of anger

According to Lakoff and Kvecses (1987), the major tenets of the CM of anger
include the following. First, there is an underlying general complex metaphor,
named the physiological effects of an emotion stand for the emotion.
The conceptualization serves as the basic structure for a large number of meta-
phoric expressions which instantiate anger based on the physiological effects that
the emotion has on a person. The conceptual principle therefore motivates the
expression of metaphors of emotion which are based in embodied experience.
Second, there are several types of conceptualized bodily experience that form
the metaphoric basis of the linguistic forms. These types include body heat,
internal pressure, skin redness, agitation, and impaired visual acuity.
Several important points about these terms need to be related here. According to
Lakoff and Kvecses, the first two types form the basis for the third type, since body
heat and internal pressure are assumed to lead to skin redness. Also, the concept of
skin redness includes specifically the face and neck of the person experiencing
anger (their original name, which included this information, has been shortened
for the purpose of brevity; impaired visual acuity is also a shortened version of
the original version, interference with accurate perception). Finally, agita-
tion specifically describes the agitation of the body, for example, when a person
shakes their fists as an expression of anger. In Lakoff and Kvecses view, these five
conceptualizations form the embodied, experiential basis for the CM of anger.

4. In Chapter 3, recent diachronic research is presented that further suggests that the differ-
ences in the two CM are the result of two different intersubjective perspectives employed to in-
terpret a scene.
Chapter 2. Diachronic aspects of synchronic concepts

body heat
Billys a hothead.
They were having a heated argument.
internal pressure
When I found out, I almost burst a blood vessel.
He almost had a hemorrhage.
skin redness
She was scarlet with rage.
He got red with anger.
I was hopping mad.
He was quivering with rage.
impaired visual acuity
I was beginning to see red.
I was so mad I couldnt see straight.

Figure 1.5 Linguistic expressions for anger in American English

Selected examples of metaphoric expressions instantiating the five basic conceptu-

alizations are shown in Figure 1, with the lexical item(s) marking the instantiated
conceptualization in italics.6
The metaphoric expressions map the source domain of physical experience
onto the target domain of anger. The effect is a verbal expression of anger which
any English speaker would immediately recognize in speaking or writing. In this
way, embodied experience is encoded in cognitive conceptualizations that in turn
motivate metaphoric expressions in language; thus, there is a cognitive link be-
tween concrete embodied experience and abstract emotion.
The expressions support the tenets of conceptual metaphor theory in two
ways. First, Lakoff and Kvecses claim that the metaphoric expressions serve as
evidence for the physiological effects of an emotion stand for the
emotion conceptualization. In addition, the expressions which encode the con-
ceptualization are highly elaborated (i.e., have related variations), to cover special
situations in communication. The elaborations provide details that support the
five basic conceptualizations of the body discussed above body heat, internal
pressure, skin redness, agitation, and impaired visual acuity.

Sub-variations of the CM of anger

The metaphoric expressions instantiate the primary CM of anger in Lakoff and

Kvecses analysis, named anger is heat. heat has two sub-CM; one for fluids as

5. Figures are used in this volume to separate data from data analysis.
6. All data are taken from Lakoff and Kvecses (1987).
Metaphor across Time and Conceptual Space

the source domain, called anger is the heat of a fluid in a container, and
one for solids as the source, named anger is fire. The first one is seen by the re-
searchers as the more basic of the two because it is more highly elaborated; it will
be discussed in detail below.

The fluid CM

The anger is the heat of a fluid in a container CM (hereafter, fluid), is

formed from two other CMs; the body is a container for emotion and anger
is heat. heat of a fluid in a container is the source domain and anger is the
target domain in metaphoric expressions. Metaphoric expressions for the resulting
fluid CM include:
You make my blood boil.
I had reached the boiling point.
These are clearly related to the CM of anger. However, Lakoff and Kvecses also
list several other metaphoric expressions, which seem to be related to a different,
unidentified CM:
Simmer down!
Let him stew.
Finally, they include a sample that is described as a historically derived instance
(p. 198) of the fluid CM:
She was seething with rage.
These particular samples clearly instantiate the anger is heat CM, but they also
appear to go beyond it. The simmer and stew samples in particular seem to instan-
tiate a cooking CM of some kind, a concept that is not within the fluid CM.7
The sample indicates the presence of a difference conceptualization of anger that
employs a source domain that is outside of the human body, within the semantic
frame of cooking.
In Lakoff s (1987) book, Women, fire, and dangerous things, the divergent na-
ture of these samples is explained by stating that Although both of these are cook-
ing terms, cooking per se plays no metaphorical role in these cases. It just happens
to be a case where there is a hot fluid in a container. This is typical of lexical elabo-
rations (Lakoff, 1987, p. 384). Lexical elaboration is the term used by Lakoff to
explain how new variants of a CM are created which cover special situations of use

7. In addition, it is unclear why a conceptualization of the human body would map a cooking
concept; one could just as easily say Cool down and remain in the human body source domain.
Chapter 2. Diachronic aspects of synchronic concepts

of the CM; replacing a word with a different one extends the metaphoric meaning,
adding new details and connections in the CM. As a result, the variations lead to
new metaphoric expressions in language.
Overall, Lakoff claims that changing lexical items in the fluid CM creates a
new elaboration and new metaphoric expressions that happen to employ cooking
terms. The purpose of these elaborations is to take advantage of semantic connec-
tions that are available in the cooking terms that are not available in other lexical
items. Hence, stew is selected in order to employ the idea of anger continuing over
an extended period of time. We will return to this issue, particularly the concept of
the cooking frame, later in this chapter.

Elaborations of the fluid CM

Elaborations of the CM (i.e., productive variations) may be created to cover special

situations. In the fluid CM, these variants include the addition of a heat scale
(or intensity) and the production of steam as a result of the heat and pressure in
the container. Examples of these elaborations are shown in Figure 2.
pressure increases eventually lead to the destruction of the container
(e.g., explosion) as a result of increasing heat, steam, and pressure (Figure 3).
However, the intense anger produces pressure on the container variant
(Figure 2, below) has two elaborations which avoid destruction of the container.
The first elaboration controls pressure by suppressing its release (Figure 4).
These two variants of pressure are interesting, especially I gave vent to my
anger (Figure 5), because vent is a verb that was commonly used in a historical
metaphoric expression of emotion in English.8 A typical form of the expression is
as follows:
I vented my spleen.

when the intensity of anger increases,the fluid rises

His pent-up anger welled up inside him.
We got a rise out of him.
My anger kept building up inside of me.
intense anger produces steam
She got all steamed up.
I was fuming.
intense anger produces pressure on the container
He was bursting with anger.
I could barely contain my rage.

Figure 2. Elaborations of the fluid CM

8. This is an archaic expression today, according to the Oxford English Dictionary Online.
Metaphor across Time and Conceptual Space

when anger becomes too intense, the person explodes

When I told him, he just exploded.
She blew up at me.
when a person explodes, parts of him go up in the air
I blew my top.
He hit the ceiling.
I went through the roof.
when a person explodes, what was inside him comes out
His anger finally came out.
Smoke was pouring out of his ears.
She was having kittens.
My mother will have a cow when I tell her.

Figure 3. Elaborations of container destruction in the fluid CM

I suppressed my anger.
He managed to keep his anger bottled up inside him.

Figure 4. Elaborations of pressure suppression in the fluid CM

anger can be let out under control

He let out his anger.
I gave vent to my anger.
Channel your anger into something constructive.
He took out his anger on me.

Figure 5. Elaborations of pressure release in the fluid CM

The question is whether the samples given by Lakoff and Kvecses for their elabo-
rations are instantiated by the hot fluid CM or by a different CM that is histori-
cally different from the present-day variants of anger is heat. Since Lakoff and
Kvecses conducted a synchronic study, they did not consider this question; their
1987 study was not designed to investigate the contribution of diachronic lan-
guage use and change to the synchronic forms that they studied.

A research gap

These facts suggest the existence of an important research gap the historical
forms should be investigated for their effects on the linguistic metaphors and,
through the relevant cultural models, the effects on the relevant CM. The study of
the historical spleen metaphor, including its relationship to Lakoff and Kvecses
CM of anger, is discussed in Chapter 7.
Chapter 2. Diachronic aspects of synchronic concepts

The experiential scene

The anger prototype scenario

The previous discussion and analysis indicates that diachronic cultural models in-
fluenced the synchronic metaphor data analyzed by Lakoff and Kvecses (1987).
This influence extends to the experiential scene proposed to instantiate the
linguistic forms. The researchers present the anger Prototype Scenario as the pro-
totypical experience that motivates the expression of anger in English-speaking
culture, conceptualized by the anger is heat CM. The basis of the scenario is the
experiential scene in which anger is expressed verbally by a person. There are five
stages in the scenario, beginning with a social event that causes a person to feel
anger and ending with a reciprocating act of revenge by the offended person
against the offending person. The five stages are shown in Figure 6.
Lakoff and Kvecses provide explanatory details for each of these stages. In
Stage 1, the offending event is an intentional act of wrongdoing by the offender,
such that the offender is guilty and the offended person has done nothing to war-
rant the offense. The offending act creates anger in the offended person. In Stage 2,
the anger continually increases on the intensity scale, and effects of the increase
are felt in the body. The effects include those discussed previously body heat,
internal pressure, and agitation of the body. The increasing intensity also leads to
an attempt to control the anger because of social norms for controlling anger and
a desire to limit the emotions physically and psychically damaging effects, both for
the angry person and others in the scene. In Stage 3, the person performs actions
to control the anger. In Stage 4, the person fails to control the anger because the
intensity has become too great; at that point, anger is expressed visibly in the body
and in the persons behavior (e.g., shaking body, facial expression, verbal expres-
sions), and also leads to a desire for revenge. Finally, in Stage 5, the angry person
takes retribution against the one who caused the offense in the first place. After the
act of retribution is completed, the anger level drops to the low end (typically zero)
of the intensity scale.
The stages are situated in a conceptualized scene in which a person proceeds
through the stages in the course of expressing anger physically and verbally. The
scene is the result of experience in the world, in which people view others (and

Stage 1: offending event

Stage 2: anger (visible to observer)
Stage 3: attempt at control
Stage 4: loss of control
Stage 5: act of retribution

Figure 6. The five stages of the anger Prototype Scenario (Lakoff & Kvecses, 1987)
Metaphor across Time and Conceptual Space

themselves) experiencing and expressing anger. After many repetitions, the scene
is schematized cognitively and is conceptualized as a series of relations between a
presupposed event that causes anger and the physical and verbal expression of the
anger. As a conceptualization, the anger scenario is profiled (i.e., in the foreground
of the conceptualization) in a base (i.e., the background domain, ICM, or frame).
The base is the presupposed, unprofiled aspects of the conceptualization, includ-
ing the event(s) that prompted the person to feel anger. The key question: what is
the backgrounded event? The question is important because the scenario cannot
be conceptualized without the base event, following Lakoff and Kvecses analysis.
However, they do not discuss the background of the scene, except in their descrip-
tion of Stage 1.

Cultural models as bases in conceptualization

Cienki (1999) argues that cultural models are part of the unprofiled base of a con-
ceptualization (p. 198) for a profiled metaphor. Understanding that anger is con-
ceptualized differently in different cultures (see later in this chapter for details), the
background event that triggers anger may vary cross-culturally, depending on the
cultural models employed in a speech community. One way to analyze the back-
ground event is via Fillmores (1982) frame construct. Goldberg (2010) states that
all words meanings are motivated by a specific semantic frame: the word diameter
profiles the line that bisects a circle, and the background frame includes the circle
(p. 40). Goldberg found in her analysis of the frames for verbs that some verbs
included distinct subevents occurring in a linear, temporal order, similar to the
anger scenario, while other verbs did not include distinct subevents. For example,
the word saut profiles the heating fat in a pan and stirring food in the fat, but the
heating and the stirring are not distinct subevents because they occur simultane-
ously ...the stirring may continue beyond the heating, but it is no longer saut-
ing once the pan is removed from the heat (p. 43). Based on this analysis, Goldberg
concludes that an (underived) verb sense constitutes a semantic frame of
predication...a generalized, possibly complex state or event that constitutes a cul-
tural unit,9 a conceptual unity of the event(s) and cultural knowledge expressed by
the verb sense. The cultural unit that constitutes the verb constrains the employ-
ment of a particular verb sense in a construction (p. 49). Studies of verb serializa-
tion also show that different languages distinguish or combine different subevents
in verbs (e.g., Bruce, 1988; Durie, 1997; Enfield, 2002). From these facts, the pre-
supposed, unprofiled aspects of the conceptualized base do not necessarily include
a causal event to explain or justify the expression of anger that is profiled in the

9. As discussed in Chapter 1, Goldberg bases the concept of cultural unit on Enfield, 2002.
Chapter 2. Diachronic aspects of synchronic concepts

conceptualization of the emotion, but if an event is present in the base, it is sup-

plied by a cultural model.
In any case, since the introspective data samples presented in Lakoff and
Kvecses (1987) study are limited to one sentence isolated from the specific com-
municative and social contexts, the analysis of the presupposed event(s) in these
samples is unclear. In Chapter 3, previous diachronic research studies will be pre-
sented to discuss the application of contextual cultural models to analyze the CM
of anger.

Atypical cases of anger

Finally, Lakoff and Kvecses state that other variants of the CM of anger are found
in English. However, in their view these variants for social situations and concep-
tual instantiations are atypical, due to differences with both the characteristics of
the CM of anger and the prototype scenario. The researchers provide a list of
twenty types of anger that have these differences. They also state that the defining
characteristic of all of the non-prototypical cases is that [t]here appear to be no
necessary and sufficient conditions that will fit all these cases (p. 217). In other
words, the non-prototypical cases vary from the prototypical cases in significant
ways, and no identifiable set of properties can define all of the non-prototypical
cases as a group.
Lakoff and Kvecses are correct concerning the lack of one set of features
(expressed in a conceptualization) that fits all of the non-prototypical cases; how-
ever, close analysis of these cases reveals several characteristics which have impor-
tant implications for the CM of anger and conceptual metaphor theory. We will
discuss Lakoff and Kvecses analysis of these cases, in order to show that the non-
prototypical cases may be systematically instantiated by another, unidentified CM,
termed the spleen metaphor. Based on the following analysis, 12 of the 20 atypical
cases were separated into two distinct (but related) groups.

controlled response over time

Of the twenty non-prototypical cases, nine appear to be related by the presence of

the same four characteristics. The cases are listed below.
Each of these cases is described by Lakoff and Kvecses, and the descriptions
are similar on several conceptual dimensions. First, unlike Stages 2 and 4 in the
anger Prototype Scenario, each case describes the control of anger as having
been successful. For example, controlled response involves staying in control of
the anger while taking non-violent, conscious retribution (for Lakoff and Kvecses,
the absence of violence apparently constitutes the presence of control, though
Metaphor across Time and Conceptual Space

Redirected anger
Controlled response
Constructive use
Successful suppression
Controlled reduction
Slow burn
Nursing a grudge
Cool anger
Cold anger

Figure 7. Non-prototypical cases of anger: controlled response over time

the discussion in their paper only implies the point). Note that the controlled
response category includes as an example the metaphoric expression He vented
his anger on her, which has similarities to I gave vent to my anger, discussed earlier
in the anger can be let out under control elaboration. The use of samples
related semantically in both the prototypical and non-prototypical elaborations
raises the question of the true status of the metaphoric expressions employing the
verb vent. This issue will be discussed further at the end of this chapter.
Second, in contrast to the CM of anger, in each case intensity either re-
mains at the same level (i.e., does not increase) or decreases as a result of control-
ling the anger. For example, in controlled reduction, anger intensity is reduced
by the conscious effort of the offended person. Third, physiological effects do not
manifest themselves visibly. An example is cool anger. In that type, the offended
person controls the anger such that physiological symptoms and effects of anger
are not visibly manifested. Finally, all of the cases exhibit the characteristics dis-
cussed above over time, instead of displaying them immediately as the result of an
offending event, as in the anger Prototype Scenario. For example, nursing a
grudge spreads control over an extended period of time; similarly, slow burn
spreads a constant level of intensity over time. In sum, each of the nine cases
exhibits the elaborations of control and intensity, with few (or zero) signs of
visible physiological effects of anger, and these factors extend over time. This
group of cases is termed controlled response over time.

intense response over time

The time factor also applies to another, smaller group of the non-prototypical cas-
es. There are an additional three cases in the list that specifically include extended
time in Lakoff and Kvecses analysis; the names of these cases are as follows.
These types extend over time in the same way that the controlled response
over time do, with intensity at a constant level, but control is no longer
Chapter 2. Diachronic aspects of synchronic concepts

Insatiable anger
Frustrated anger

Figure 8. Non-prototypical cases of anger: intense response over time

present as the person is performing acts of retribution over the extended period of
time (though visible anger is not necessarily present, as in the controlled
response group). As an example, Lakoff and Kvecses (1987) state that for insa-
tiable anger, the intensity of the anger stays above zero and the anger continues
to exist (p. 214). Similarly, for frustration and wrath, the intensity also re-
mains above zero and extends over time; Lakoff and Kvecses characterize wrath
this way: The intensity of the offense is very great and many acts of retribution are
required in order to create balance. The intensity of the anger is well above the
limit and the anger lasts a long time (p. 216). With all three of these cases, anger
is at such a high level of intensity that repeated acts of retribution over an extended
period of time are required to reduce the intensity level. Due to the significant
elaborations of intensity and time, this group of three cases is termed intense
response over time.
The most important issue in this analysis is that the experiential scene that
motivates these 12 non-prototypical CM of anger appears to be different from the
one detailed in Lakoff and Kvecses Anger Prototype Scenario. Comparing the
two scenes on the basis of the stages of the original Scenario (renamed the Blood
Anger Prototype Scenario for the purposes of this comparison), the differences
are more clearly delineated. The differences between the two Scenarios are marked
by asterisks (*) in the Spleen Anger Prototype Scenario column.
From this comparison, the two scenarios appear to be motivated by different
conceptualizations of anger. If the sub-events that motivate the conceptualization
are not the same, then it follows that the scene is viewed in each Scenario from
specific and also different perspectives. Further, the presence of different perspec-
tives indicates that different cultural models provide each perspective. These con-
clusions were investigated in the studies detailed in this volume.

Blood anger Prototype Scenario Spleen anger Prototype Scenario

Stage 1: offending event (presupposed) offending event (presupposed)

Stage 2: anger (visible to observer) anger (not visible to observer)
Stage 3: attempt at control attempt at control
Stage 4: loss of control maintains control
Stage 5: act of retribution (invariable) act of retribution (variable)

Figure 9. Two experiential scenes of anger: A comparison

Metaphor across Time and Conceptual Space


The analysis of Lakoff and Kvecses 1987 data on the non-prototypical cases of
anger indicates that time, intensity, and the lack of physiological effects apply
to 12 of the 20 atypical cases in their list; control was found in nine of the
12 cases. The dimension of control is not present in the intense response over
time group due to the presence of acts of retribution. Interestingly, this explains
acts of retribution when physiological effects of anger are not present. Therefore,
for all 12 cases, control is defined as having the property of manipulation, al-
lowing conscious, premeditated retribution, both violent and non-violent. In sum,
the same four factors time, intensity, lack of physiological effects, and control
are present and/or manipulated in all 12 cases. The analysis indicates that there
may be a systematic relationship between the four factors in the experiential scene.
The question is whether the atypical cases are unrelated to each other (i.e., the
similarities are coincidental) or indicate the presence of a different CM than Lakoff
and Kvecses CM of anger. The analysis has provided some evidence to support
the existence of a second CM, including the diachronic influence of the verb vent,
the presence of the cooking frame in some samples, the two anger scenarios,
and the differences in perspective on the experiential scene which indicate differ-
ent cultural models. Additional evidence is needed to show that these are not co-
incidental artifacts of the data and/or the analysis.

Classification of the non-prototypical cases

At this point, the key question is the following: Does the evidence discussed above
justify classifying these cases as minor variants of the heat CM (as Lakoff and
Kvecses argue), major dimensions of the domain of anger, or a different CM
altogether? The question turns on whether the atypical cases are motivated system-
atically by a different experiential scene and/or a different cultural model, which
would in turn provide a different intersubjective perspective, conceptual domains,
and domain mappings; if so, then they are likely instantiated by a different CM.
The fact that the atypical cases do not enact Stages 2, 4, or 5 in Lakoff and Kvecses
anger Prototype Scenario suggests that the experiential scene for the non-proto-
typical cases is different from the hot fluid metaphors in Lakoff and Kvecses
(1987) analysis. If the experiential scene itself is different, then the conceptualiza-
tion that is produced by the scene is different, as well.

control in the non-prototypical cases

Another important issue is the ability of the person to control the anger when the
anger Prototype Scenario assumes that this is not possible. In their discussion of
Chapter 2. Diachronic aspects of synchronic concepts

their 1987 proposal for the anger Prototype Scenario, Lakoff and Kvecses state
The course of anger depicted in the prototype scenario is by no means the only
course anger can take. In claiming that the scenario is prototypical we are claiming
that according to our cultural theory of anger, this is the normal course for anger
to take. Deviations of many kinds are both recognized as existing and recognized
as being noteworthy and not the norm (pp. 211212 [italics added]).

They are right to state that the non-prototypical cases are different from the hot
anger CM, and evidence is given to show that the scenario is the expected re-
sponse when a person experiences feelings of anger. However, their comments do
not specifically address why the non-prototypical cases are in fact not the norm.
For example, the application of the control feature in Stages 2 and 4 in the
non-prototypical cases appears similar to the application of deictic orientation in
experiential scenes: the presence of control is possibly a result of a cultural mod-
el selecting the control feature from among various dimensions in the CM of
anger. In both Stages 2 and 4, the person is able to control the anger, suppressing
visible physical effects in many cases and maintaining control and not engaging
in retribution. The answer is unclear at this point, but the parallels to Heines (1997)
description of deictic orientation are interesting. It is possible that a different cul-
tural model is selecting control in the non-prototypical cases. If this analysis is
correct, then the Spleen anger Prototype Scenario may have a different historical
and cultural origin.

Theory as a test of observed language patterns

Finally, the systematic application of an experiential scene is an important issue in

the study of CM for another reason. CF theorists have claimed that the systematic
patterns employed in language are evidence of cognitive processes. Deignan (2006)
points out that Lakoff was persuaded that metaphor is central to abstract thought
when he found that there were systematic relations that linked different linguistic
forms; other CM researchers make similar claims (p. 107). Lakoff s conclusion is
of great theoretical importance to understanding the relationship between lan-
guage and cognition. Yet, studying language to develop theory is not enough.
Deignan concludes that,
Given this importance placed on language as evidence for the theory, it does not
seem unreasonable for a descriptive linguist to turn the relationship around: to
look to the theory for a possible account of the patterns that he or she observes in
naturally-occurring language (p. 108).
Metaphor across Time and Conceptual Space

A basic principle of scientific inquiry implied in the above statement is, Theories
have to be tested in order to understand the relationship between abstract thought
(cognition) and linguistic metaphor in real-world language use (Steen, 1994, p. 9).
The current study has adopted Deignans descriptive linguistic view that theory
should be applied to data, to see if the proposed mode describes the data accu-
rately. If so, then the theory is useful for describing the relationship between lan-
guage and cognitive processes; if not, the theory may be in need of revision to fit
the known facts.

The proposal: The blood and spleen metaphors

This analysis of Lakoff and Kvecses (1987) CM of anger indicates that more
detailed research is needed to investigate the specific conceptualizations, includ-
ing cultural models, that motivate the two different linguistic metaphors, His blood
boiled, identified by Lakoff and Kvecses as motivated by the prototypical CM of
anger, and He vented his spleen, identified as a non-prototypical case of the same
CM. Understanding these potential differences, in the current studies the two
metaphors are given separate terms: the first linguistic metaphor is termed the
blood metaphor (see also Sim, 2011), and the second is termed the spleen meta-
phor. These two linguistic metaphors will form the basis for the investigation of
diachronic metaphors of anger. The next section reviews synchronic research
studies that have investigated the CM of anger and cultural models.

Synchronic studies of CM and culture

The purpose of this section is to review studies of CM in present-day, synchronic

research, in light of Lakoff and Kvecses (1987) analysis of the CM of anger.
Though the current studies in this volume employ a diachronic design, there is an
important reason for analyzing synchronic linguistic forms: to delineate the issues
that should be addressed in studying the separate roles of bodily experience and
cultural knowledge in forming cognitive concepts. The section reviews studies of
bodily experience that also found important influences on conceptualization from
cultural knowledge.

Conceptual metaphors as universals

Matsuki (1995)

Several synchronic studies which were intended to study bodily experience

found evidence that cultural knowledge was an important component of the
Chapter 2. Diachronic aspects of synchronic concepts

conceptualization. For example, Matsuki (1995), studying anger metaphors in

Japanese, found that the metaphoric expressions employed the same embodied
conceptual metaphor described by Lakoff and Kvecses anger is heat. In addi-
tion, the study found specific evidence for anger is a hot fluid in a container,
including the container image schema, heat, pressure on the fluid, and visible
physiological effects (skin redness, bodily agitation, and interference with visual
perception). Yet, Matsuki also noted some differences; for example, the container
in the Japanese metaphors is the stomach (belly), not the human body; also, the
Japanese word for belly (hara) is used when anger rises to the head (atama); the
researcher states that the substance that comprises hara is unclear since the stom-
ach cannot rise physically to the head. Finally, Matsuki found differences between
the anger prototype scenario described by Lakoff and Kvecses and the scenario
found in Japanese; native-speaking Japanese informants stated that they would not
lose control of their anger as the scenario dictates (a similar result was found in the
atypical metaphors in the Lakoff and Kvecses (1987) study; see the discussion
earlier in this chapter). Matsuki stated that the differences in Japanese are the re-
sult of individual idiosyncrasies (1995, p. 149); he concluded that the American
English conceptualization of anger is partially applicable to Japanese anger
(p. 150). The Japanese conceptual metaphor of anger exhibits the characteristics
of universality and intersubjectivity that Lakoff and Kvecses found for American
anger CM and also characteristics of shared cultural knowledge.

Yu (1995)

Yu (1995), discussed briefly in the previous chapter, presented results for Chinese
metaphors of anger which came to similar conclusions some aspects of the
samples follow the Lakoff and Kvecses model, and other aspects incorporate
Chinese cultural knowledge, particularly medical practices. For example, though
the metaphoric expressions of anger in Chinese conceptualized heat, pressure,
and visible physiological effects, gas was instantiated, rather than the fluid found
in the English metaphors. In addition, the Chinese metaphors employed more in-
ternal organs than the English metaphors. Yu attributes the use of gas in the CM to
the physical properties of gas; the use of internal organs is explained by cultural
beliefs, specifically traditional Chinese medical practices. As a result, The under-
lying cognitive model based on the fundamental theories of Chinese medicine has
led to a cultural emphasis in China of sensitivity to the physiological effects of
emotions on the internal organs. This, in turn, has influenced the way Chinese
people talk about emotions (Yu, 1995, p. 85). This result is similar to Geeraerts
and Grondelaers (1995) study, which found evidence for the influence of Renais-
sance-period medical beliefs on the conceptualization of anger in English and
Metaphor across Time and Conceptual Space

Dutch (see Chapter 3). Despite this finding for the effect of cultural knowledge on
conceptualization, Yu concludes that the Chinese anger metaphors reflect the
same universal embodied experience found in English.

Conceptual metaphors as cultural models

Maalej (2004)

Other studies also found evidence for cultural beliefs within the CM of anger,
including entire CMs based on cultural knowledge. Maalej (2004) investigated
Tunisian Arabic. The result of the analysis found three types of embodiment:
(1) physiological embodiment; (2) culturally tainted embodiment; and, (3) cultur-
ally specific embodiment (hereafter, CSE). The first type is the same as Lakoff and
Johnsons embodied realism; the other two are variants which include increasing
influence from shared cultural knowledge, with culturally specific embodiment
displaying the most influence from culture. Maalej concludes from the analysis
that the embodiment principle needs to be broadened to include the two new vari-
ants, allowing non-human forms of embodiment (e.g., a sheeps stomach as a
source domain mapped onto human anger, the target domain) and inanimate
forms of embodiment (e.g., a dust storm mapped onto human anger). This formu-
lation is a significant departure from Lakoff and Johnsons original construct.
Cultural Models vs. the Experiential Scene. The problem inherent in Maalejs
proposal for broadening the embodiment principle is that his definition of em-
bodiment is not grounded in the human experience of the experiential scene. In
particular, culturally specific embodiment (CSE) allows for metaphors which in-
stantiate non-human forms of embodied experience of a scene (Maalej, 2004,
pp. 6667). All of the CSE examples provided in the study map a source domain of
a non-human physiological body part (e.g., a sheeps stomach) or an inanimate
entity (e.g., a dust storm) onto a target domain of a human emotion. The human
physiological basis of experience of the scene is completely detached from the em-
bodiment principle, decoupling CMs from their empirical grounding in human
experience. Calling the CSE mapping a form of embodiment expands the concep-
tual metaphor construct to the point that it has little meaning.
Constraints on the Mapping of Conceptual Metaphor. An extended discussion
of the problem is found in Lakoff and Turner (1989). In their book, More than cool
reason: A field guide to poetic metaphor, the authors identify factors that constrain
the mapping of a cognitive concept to another, and the question Can a metaphor
exist between any two things? is specifically discussed. The answer is no: But this
phenomenon our wide-ranging ability to find ways to metaphorically link two
linguistic expressions does not mean that metaphor is completely unconstrained,
Chapter 2. Diachronic aspects of synchronic concepts

that anything can map onto anything any old way (p. 200). One example to sup-
port the researchers contention is the metaphoric expression Death is a magician;
Lakoff and Turner show that magician performs an action which causes death, but
magician does not map onto the dying persons last breath, an action which is the
result of death, not the cause. After analyzing several examples, the researchers
conclude that [t]hough wide-ranging metaphorical interpretations are possible,
they are far from arbitrary...[i]t is not the case that anything maps onto anything
(p. 203). Conceptual metaphor is based on the experience of a scene in the human
body a non-embodied linguistic metaphor is metaphorical, but it is not moti-
vated by a CM.
If embodied realism is broadened to include non-human source domains like
a sheeps stomach and a dust storm (neither of which can be experienced physio-
logically by a human), then embodiment as a construct can theoretically include
any experiential scene from the perspective of any concrete physical object an
airplane or the Moon (Lakoff & Kvecses, 1987). Conceptual metaphor would be
extended to everything and anything in an arbitrary fashion, undermining the
theory as a principled account of how a certain source domain is mapped with a
certain target domain. Ultimately, if the theory cannot explain why a particular
mapping occurs, then its usefulness in research is effectively nil. This issue also
points to the effective limits of any theory no theoretical principle explains every
instance of a phenomenon. CSE is not the result of construal concerning human
experience of a scene in the world, and therefore these types of metaphor can be
eliminated from consideration as CM. The Maalej (2004) study is an important
one for delineating the limits of culture on conceptualization. Conceptual meta-
phor theory must limit the perspective of an experiential scene to human experi-
ence of the world, in order for the theory to have empirical grounding.

Kvecses (2010a)

Kvecses (2010a) comes to conclusions similar to the Maalej (2004) study but pro-
poses a somewhat different motivation for non-embodied metaphor. In a study of
metaphorical creativity, a new form of linguistic metaphor is proposed, termed
context-induced contextual information, such as cultural models, motivate the
metaphor; similar to CSE, universal embodiment does not play a role. This new
form is proposed to account for the novel use of metaphor in creative works, such
as poetry.
Kvecses argues that context-induced metaphor is a new sub-class of resem-
blance metaphor. This view assumes a sufficiently broad definition of resemblance
to cover both embodied conceptual metaphor and non-embodied linguistic meta-
phor. According to Grady (2008), there are two types of resemblance metaphors:
Metaphor across Time and Conceptual Space

image/attribute and relational. The first type involves shared perceptual properties,
such as physical similarity (e.g., Ed is a beanstalk.), and the second involves shared
non-perceptual features (e.g., A cigarette is a time bomb).10 Grady states that, due
to the infinite number of possible shared properties available in non-autonomous
knowledge, the number of potential resemblance metaphors is also limitless
(p. 341). The resemblance metaphor construct as defined by Grady covers both
embodied conceptual metaphor and non-embodied linguistic metaphor, and
Kvecses proposal for context-induced metaphor fits into the latter type.
Kvecses acknowledges that his proposal has limitations as it relates to con-
ceptual metaphor theory: To some, however, to say that such metaphors repre-
sent a new class may be overstating the results of this study. It may be suggested
that while there is not always a bodily basis, there is always some resemblance on
which metaphors are based (Kvecses 2010a, p. 692). The resemblance therefore
could be a general cognitive process, such as analogy, rather than a mapping be-
tween source and target domains. However, the point made previously concerning
non-embodied metaphor CSE, context-induced or a similar type still holds: a
linguistic expression can be metaphorical, but that does not necessarily indicate
that the metaphor is motivated by human embodied perspective of an experiential

Conclusion: embodiment and cultural models as equal partners

A variety of other research has concluded that a CM is instantiated by a complex

mix of multiple factors, including embodiment and cultural models. Kvecses
(2005; 2009) pressure of coherence model, discussed in Chapter 1, is one example.
Another is Barcelona and Soriano (2004), which studied anger metaphors in
Spanish and English. The authors found eight metaphors of anger that both lan-
guages share. There were some differences; Spanish does not conceptualize steam-
ing as a physical effect of anger (thus, a language-specific sub-mapping), and
similarly English does not instantiate frying (p. 301). The paper recommends a
multidisciplinary approach to research, employing cultural, neural, psychological
and linguistic accounts (p. 307) in order to better understand the interaction be-
tween language and cognition. MacArthur (2005) collected metaphoric expres-
sions of horse riding, a human embodied activity. The author argues that common

10. The two types and the example sentences are from Grady, 2008, p. 341.
11. An alternative construct (Cameron 2008) is systematic metaphor, applied specifically to the
systematic use of source domain concept(s) in a study of spoken discourse. This type describes
the mental representations of individuals (p. 323) rather than intersubjective conceptualiza-
tions of human experience.
Chapter 2. Diachronic aspects of synchronic concepts

experience, whether directly experiential or vicarious, does not account for the
high prevalence of horse riding metaphoric expressions; rather, the metaphor was
spread by the upper classes of society, who were the primary horse riders and also
influential in setting social trends. Social transmission and propagation of the con-
ceptualization is the result of both experiential and social factors. Similar to
Barcelona and Soriano, MacArthur recommends taking into account multiple fac-
tors when analyzing CMs. Cienki (1999), in a study of two Russian words for hon-
esty, concluded that the different conceptualizations found in each word are
possibly evidence of general patterns in Russian culture which organize or link
up families of related cultural models and so can provide coherence to a shared
worldview (p. 200). Cultural models provide organizing principles and associa-
tions between concepts to foster cognitive structuring in the mind. Again, like the
authors discussed previously, Cienki suggests studying CM in a variety of ways,
including the study of linguistic metaphors in context, to determine the full range
of cultural influences on metaphor instantiation. All of these studies accept the
influence of culture in CM as an important factor along with embodiment and all
recommend multidisciplinary research designs.

Frequency of use and cultural models

The frequency of use of a metaphor is also affected by cultural knowledge. Sim

(2011) studied linguistic metaphors in English and Hungarian that employed
blood as a lexical item. One of the results showed that in cases in which blood col-
located with boil in English, 75% of the samples instantiated anger. However, in
Hungarian, only 41% of the cases denoted anger; the other 59% expressed excite-
ment or vitality (p. 2907). In addition, the number of instances varied by genre: in
the American cases, the majority were found in non-fiction texts, whereas in
Hungarian, more cases were found in fiction. The researcher concludes, support-
ing an earlier study that investigated the blood metaphor (Mischler, 2008), that a
specific metaphorical expression can signal different emotions in different lan-
guages. In light of these findings, Sim states that more diachronic and discourse
analysis-style investigations of a number of expressions is necessary (p. 2909).
Corpus studies in particular are recommended, which can delineate fine details of
variation in conceptualizations within and across languages and genres.
Finally, Emanatian (1999) takes Cienkis (1999) analysis of families of cultural
models a bit further, suggesting an embodiment/culture continuum or scale. The
study investigated Chagga, a Bantu language, and found that cultural models are
an important, though highly variable, factor in metaphors of sex and eating. The
study concludes that the separate influences of embodiment and culture vary in
their significance from CM to CM; other studies assume that one of the two factors
Metaphor across Time and Conceptual Space

dominates the CM under study. As a result of the analysis, Emanatian suggests that
each CM should be studied separately as a unique instantiation of multiple cogni-
tive, cultural, and linguistic factors. The Emanation study shows that the
relationship between embodied experience and cultural knowledge is complex,
and describing both is important to understanding the content and meaning of a
conceptual metaphor.

Chapter summary and conclusions

Overall, the synchronic studies conclude that conceptual metaphor is motivated

by bodily experience, yet cultural knowledge was found to be an important factor
in motivating the metaphoric expressions in each case. The analysis of Lakoff and
Kvecses (1987) data similarly indicates that cultural and historical factors may
influence the CM of anger, as well as the atypical cases analyzed in their study.
Moreover, the research literature tends to follow a theoretical assumption that
universal metaphors are possible (e.g., Kvecses (2005) argument for potentially
universal metaphors), but other research studies continue to find cultural knowl-
edge inextricably entwined with embodiment, as the studies by Cienki (1999),
Emanatian (1999), Kvecses (2010), Maalej (2004), Matsuki (1995), Sim (2011),
and Yu (1995) found. The relationship between embodied experience and cultural
knowledge in conceptualization requires more study, due to the differences be-
tween current theory and research results.
Finally, several of the studies recommend multidisciplinary research designs
in order to more fully capture the influence of cultural models on CMs, including
the use of text corpora to collect a larger number and wider variety of samples.
Barcelona and Soriano (2004), Cienki (1999), MacArthur (2005), and Sim (2011)
all recommend further studies with these design characteristics. In particular,
MacArthurs study concerning the historical influence of social groups on meta-
phor spread in society12 echoes Sweetsers (1990) assertion that diachronic culture
and synchronic language are connected in tangible ways. However, cultural and
historical influences are difficult to study with present-day data, as Chapter 1 dis-
cussed. Multidisciplinary, longitudinal, and empirical study designs are needed to
delineate the complex relationship between bodily experience in the world and
cultural models.

12. Chapter 3 will present diachronic studies that also support MacArthurs conclusion.
Chapter 2. Diachronic aspects of synchronic concepts

Filling the research gap

The studies described in this volume filled the research gap discussed above by
studying the CM of anger in a diachronic, longitudinal, and mixed methods de-
sign (i.e., includes both quantitative and qualitative analyses) to investigate wheth-
er a historical cultural model, such as the Four Humors, systematically instantiates
the complex CM; whether the cultural model is associated with changes in the
conceptualization of the CM and, how the changes affect variation in historical
and present-day metaphoric expressions. Moreover, the current studies have ad-
opted Deignans (2006) assertion that theories concerning the relationship be-
tween cognition and language should be applied to natural language use data, in
order to determine if the proposed theoretical model describes the experience of
actual language speakers accurately. If the question is answered affirmatively, then
the theory is useful for describing the relationship between language and cognitive
processes; if not, the theory may be in need of revision to fit the known facts.
In Chapter 3, previous studies of conceptual metaphor in diachronic data are
discussed in detail. Based on the conclusions of that chapter, the research design
for investigating the relationship between conceptualization and cultural models
in the historical CM of anger is described in Chapter 4.
chapter 3

Metaphor across historical time


Historical (i.e., diachronic) language forms and the cultural models that produced
those forms have important effects on synchronic iterations of linguistic expres-
sions, as Sweetser (1990) has pointed out. Bybee (1988) also argues that syn-
chronic states must be understood in terms of the set of factors that create them.
That is, we must look to the diachronic dimension... (p. 351). Changes in mean-
ings over time have synchronic effects; as a result, the current meaning of an ex-
pression can reflect accumulated historical changes over time. Examples of the
diachronic effects on synchronic linguistic expressions include the synchronic
samples from Lakoff and Kvecses 1987 study of the CM of anger, discussed in
Chapter 2. As another example, the English word foot has several current mean-
ings; for example, the term can refer to the human body part and also to a histori-
cally more recent metaphorical meaning that references the part of an inanimate
object which touches the ground and supports the object. Expressions for the sec-
ond meaning include the foot of the bed and the foot of the mountain. The addi-
tional meaning indicates a change in the cognitive conceptualization of foot; the
concept has been extended to include inanimate objects that share the human
foots conceptual entailment of support. These examples indicate the long-term
effects that diachronic cultural models have on synchronic forms, and by exten-
sion, on conceptualization.
Concerning conceptualization processes, historical changes in meaning may
indicate changes in perspective on the experiential scene, as a result of new experi-
ences and/or the effect of changing cultural models in the speech community. Re-
search on this type of historical change has important implications for CM theory
and for cognitive-functional linguistics: if changing cultural beliefs in turn change
conceptualization, then it follows that cultural models work in concert with cogni-
tion in producing construals of experience. The purpose of this chapter is to inves-
tigate previous research in diachronic conceptual metaphor. The overall goal is to
develop the research questions and study design for the current studies; the ques-
tions and design are detailed in Chapter 4.
Metaphor across Time and Conceptual Space

Two types of historical study

Synchronic-historical research

The effects of diachronic cultural models on conceptualization processes are a re-

cent line of research in cognitive-functionalism. The clear majority of metaphor
studies in this line have been designed as synchronic-historical or point-in-time
investigations. In such studies, the influence of culture is more difficult to discern
because culture can take many years to have any appreciable effect on language.
Present-day metaphoric expressions (i.e., linguistic expressions with metaphoric
meanings, such as His blood boiled, discussed in Chapter 2) reflect the influence of
historical cultural ideas that are no longer shared among speakers in the speech
community, increasing the complexity of linguistic analysis. Though these older
cultural ideas are no longer consciously acknowledged, their previous influence is
still present in the structure and meaning of the present-day language linguistic
form (see Bybee, 2001, 2003). Therefore, historical time as a variable should be
taken into account to provide a full analysis of a CM. Such study is uniquely posi-
tioned to reveal historical cultural models and their influence on present-day lan-
guage form and meaning.

Synchronic-historical research: An example

Among the synchronic-historical studies of CMs (see next section for detailed
discussion), Bertuol (2001) is a typical example. The study applies conceptual met-
aphor theory to the use of mathematical language in the poem, The circle of the
brain cannot be squared, by Margaret Cavendish (1653). As a result of the analysis,
the researcher discusses the poems central CM, universe is mathematics.
Bertuol concludes that mathematical concepts were highly influential in English
thought, culture, and language of the period.
Discussions of cultural knowledge and its influence on language in studies like
Bertuols are a natural result of historical analysis because the data reveal more
clearly the cultural knowledge that is no longer in force in present-day thought
and language. However, such discussions assume that the contrast between the
historical period and the present day is sufficient in itself to show a change in cul-
tural knowledge over time. The problem with this assumption is that the sample
(the poem) does not actually show dynamic change in culture over time, but mere-
ly the static absence of synchronic (present-day) cultural knowledge. The method
clearly shows a contrast in cultural models between the two time periods, but
change in the model over time cannot be shown because data from intervening
time periods are not analyzed.
Chapter 3. Metaphor across historical time

Though the study is valid as a detailed investigation of a specific CM, the re-
search design is synchronic for the historical time period, affording only a
snapshot view of the time-bound culture in which the CM is situated. The com-
pression of the time factor into a single poem published in 1653 obscures the ef-
fects of experience and culture over many decades which may have led up to the
instantiation of the CM. Thus, in Bertuols study design, general historical research
was conflated with diachronic research these research designs were viewed as
interchangeable when in fact they are distinct.
Other synchronic-historical studies of CM have also been conducted (e.g.,
Csbi, 2001; Goldwasser, 2005; Slingerland, 2004; Wiseman, 2007). However, as
the discussion of the Bertuol study showed, synchronic study of historical lan-
guage data collapses the culture variable into a point-in-time that does not pro-
vide information on the changes in form and meaning over time that produced the
linguistic samples under study.
To summarize, historical research, as a general category, covers any consider-
ation of the past, including synchronic study designs for the point-in-time period
under study; this type of study is termed here synchronic-historical. In contrast,
diachronic research specifically denotes longitudinal, across time studies.
Diachronic studies have an advantage to delineate the changes in the form and
meaning of linguistic expressions that result from the slow, incremental shifts in
cultural models over many years (Bybee, 2003).
Understanding these two types of historical language study and their specific
features, synchronic-historical studies, such as those discussed above, will not be
reviewed here. Instead, in the next section, diachronic studies which investigated
changes in conceptualization over time are the focus.

A synchronic-historical study of CM and culture

Geeraerts and Grondelaers (1995) is the historical counterpart for Western

European languages to the Yu (1995) synchronic study for Chinese.1 The research-
ers conducted an exploratory study in English and Dutch, collecting non-linguistic
historical data from art and medicine to investigate the effects of diachronic cul-
tural models on synchronic language use. The study concluded that historical cul-
tural beliefs in medical treatment called the Four Humors model may have had an
important influence on metaphor instantiation in the two languages. Thus, in find-
ings that parallel Yu (1995), cultural beliefs and practices concerning the human
body were important sources for instantiating the metaphoric expressions.

1. Geeraerts and Grondelaers (1995), though it is a synchronic-historical study design, is re-

viewed here due to important contributions to theory and methodology for the current study.
Metaphor across Time and Conceptual Space

In addition, the study also provided two major reasons why the Four Humors
explanation should be accepted: first, fluid as the motivator for anger is a more
parsimonious explanation when the humoral account is considered. For example,
in metaphoric expressions such as he was filled with joy and she could not contain
her joy, Geeraerts and Grondelaers question how the expression combines with
anger is heat to yield a fluid; a solid or gas which fills the container is also a
logical possibility. The humoral model does have the fluid property, and the mod-
el may provide fluid for the metaphoric expressions.
Second, the humoral account makes better sense of the samples which do not
seem to have a physiological basis. As an example, Geeraerts and Grondelaers cite
as one case that feelings of love are associated with heat in the Four Humors, but
physiological heat is not associated with sexual desire; example metaphoric ex-
pressions include burning devotion and warm feelings. Yet, these expressions do
not mean that the person feels hot: is physiologically unlikely that persons
in love have a permanently raised skin temperature... (p. 168). The implication is
that the Four Humors model explains the presence of the heat property in cases
where embodied physiological effects are absent. Geeraerts and Grondelaers
(1995) conclude that the anger is a hot fluid in a container is one of the
traces on our emotional vocabulary (p. 176) left by the humoral theory. The
metaphor therefore is not motivated directly by the physiological effects of anger,
but it is part of the historical (and reinterpreted) legacy of the humoral theory
(p. 176). Conceptual metaphors are motivated primarily by culture, not by embod-
ied experience. They conclude that a large extent, the synchronic polysemy of
lexical items is a reflection of their diachronic development (p. 177). The research-
ers note the limitations of the study, which included its exploratory scope and
synchronic design. Further studies were recommended, particularly in diachronic
designs, to test the study findings.
It must be pointed out at this juncture that the view of these researchers has
not changed since the study was published in 1995. For example, Geeraerts (2010)
reviewed the 1995 study in a discussion of the influence of culture on word mean-
ing. Based on the existence of lexical relics of the Four Humors currently still in
use in English, such as phlegmatic and bilious, he concluded that ...the purely
physiological interpretation put forward by Lakoff and Kvecses needs to be inter-
preted along cultural lines (p. 251252). The quote acknowledges the role of both
physiology and culture in conceptual metaphor, though the full extent of the con-
tributions of each is still a matter of inquiry.
In a study of anger in various languages, Kvecses (1995) countered Geeraerts
and Grondelaers conclusions that privilege culture over and above other factors,
arguing instead in favor of embodiment as the primary influence:
Chapter 3. Metaphor across historical time

If the conceptualization of anger were only a matter of culture, we would have

to have radically different conceptualizations in the case of radically different cul-
tures. The conceptions of anger we have looked at indicate this is not the case...
Thus the conceptualization of anger must be influenced by factors over and above
the particular historical development of the culture (p. 194).

Kvecses is correct his work and many others in CF research over the past
30 years show the important effects that embodied experience has on conceptual-
ization. However, this position differs from Geeraerts and Grondelaers stance
only in degree, not in kind. The two positions are in agreement that embodiment
and culture both play important roles in conceptualization processes; they differ
principally on the level of influence that each factor contributes.


The position of the current studies, described in detail in Chapter 1, is that both
embodiment and culture contribute equally, or at the least, both are influential
enough that the differences between them are largely inconsequential. Previous
study discussed in this volume so far indicates that both factors work together in
conceptualization processes; based on that work, the current studies accept this

Diachronic studies of conceptual metaphor

The research study of English metaphor described above Geeraerts and Gronde-
laers (1995) was historical in design but not specifically diachronic; the study did
not investigate change in conceptualization over longitudinal time.2 Several other
studies have employed diachronic designs investigating a single variable,3 and
these are discussed below. Following the section, a second section of multidisci-
plinary diachronic study designs is described.

2. Kvecses (1995) was synchronic and cross-linguistic, so it does not apply to the current
discussion of diachronic research in metaphor.
3. One study not discussed here, Ghesquire & Vandevelde, 2011, though not a study in con-
ceptual metaphor, deserves mention for its diachronic corpus design over an extended time scale
(approximately 900 years). The authors demonstrate the value of employing corpora for the
study of diachronic changes in the intersubjective meanings of English such and Dutch zulk.
Metaphor across Time and Conceptual Space

Diachronic studies: Single factor designs

Gevaert (2002)

Gevaert (2002) reconstructed the historical conceptual domain of anger in Old

English in a longitudinal study design which employed a frequency analysis of
words denoting anger. The data, collected from the Toronto corpus, were catego-
rized into three historical periods. The three periods are Before 850 A.D., 850950,
and 9501050; a later, additional analysis investigated briefly the conceptual do-
main of anger in Middle English, from 12001450. The three periods were cho-
sen for three reasons: (1) to follow the periods used in the Helsinki corpus (which
have status as a research standard, according to Gevaert); (2) to spread the data
more evenly for analysis; (3) and to account for cultural evolution (p. 285), by
which Gevaert refers to the most dominant culture of a historical period. For ex-
ample, the first period (Before 850 A.D.) was dominated by the old Germanic cul-
ture, and the second period (850950) was primarily influenced by Latinate cul-
ture. Each period was influenced by a unique set of historical and cultural factors,
and the three periods were devised to isolate their separate effects.
The analysis shows that the historical periods were marked by some fluctua-
tions in the frequency (called tokens) and number of different words (called types)
of anger words related to heat. In the first historical period, two words (i.e., hatheo-
rt, hygewlm) comprised 1.58% of all words for anger; in the second period, the
tokens increased to 12.81%, and the types also increased, from two in the first pe-
riod to seven in the second period (e.g., hatheort, hathige, blse, ghyrstan; hy-
gewlm from the first period was not found). Gevaert concludes that in the second
period, ...the heat-domain gains importance spectacularly due to Latin (and
biblical) influence (p. 293), an indication of the effect of cultural knowledge on
the anger is heat conceptualization.
In the third period (A.D. 9501050), the tokens decreased to 6.23%, and types
decreased to five (hathige and blse from the second period were not found; onten-
dan replaced ghyrstan). Interestingly, in the Middle English period, the word anger
first appeared, and sharp was added as a new conceptualization, which Gevaert
says fits in nicely with the Old English conceptualization of anger as something
which hurts... (p. 293). In addition, the years from 1350 to 1450 show a significant
increase in French loan words, especially those introducing new cognitive con-
cepts; words added to the English lexicon included choleric, melancolie, and boilen,
all of which are related to the Four Humors model. Gevaert concludes that the
conceptualization of anger is generally stable, but the Middle English period was
characterized by ...drastic change, apparently under the influence of the humoural
theory (Gevaert, 2002, p. 294). The study therefore indicates, like Yu (1995) and
Chapter 3. Metaphor across historical time

Geeraerts and Grondelaers (1995), that cultural beliefs concerning the human
body and medical practice (in this case, the Four Humors model) influenced sig-
nificant changes in the conceptualization of anger during the late Middle Ages.
Kvecses (2005) Response to Geveart (2002). Gevaerts specific finding for fluc-
tuations in the frequency of the heat conceptualization over time are noteworthy
in light of current conceptual metaphor theory. Kvecses (2005) summarizes the
This is an extremely important finding because it bears directly on the issue of
universality of metaphorical conceptualization across time. If the conceptualiza-
tion of anger in terms of heat is a mechanical and or automatic consequence of our
real physiological processes in anger, this fluctuation should not occur. It cannot
be the case that peoples physiological characteristics change in anger every 100 or
200 years or so (p. 105).

Kvecses rightly points out that bodily experience does not change over time;
therefore, another factor (or factors) is influencing the diachronic changes in the
heat conceptualization. In a discussion of the causes of variation in CMs, Kvecses
(2005) provides an explanation of the reasons for the changes found in the Gevaert
study. I believe the answer is that universal physiological features provide only a
potential basis for metaphorical conceptualization without mechanically con-
straining what the specific metaphors for anger will be (p. 248). The universal
potential, by implication, can be selectively instantiated in a specific CM, provid-
ing conceptual space for culture to have a role in conceptualization.
How far does cultures role extend? Can a CM consist solely of cultural knowl-
edge? Kvecses provides an answer a few pages later: As a matter of fact, it also
seems possible that universal physical or biological embodiment is entirely ig-
nored in conceptualization (p. 251). To support this statement, he cites Lutz
(1988) who analyzed song, the Ifaluk word for anger. Kvecses states that the con-
ceptualization of anger in the word did not include any of the characteristics of
the anger is a hot fluid in a container CM. Instead, song was conceptualized
by its social aspects, especially concerning how anger is resolved in social situa-
tions. Kvecses commented on Lutzs analysis as follows.
Although the Ifaluk may well have very similar physiological processes in anger to
the English and Chinese, this fact does not necessarily lead them to conceptualize
song as pressure in a container...Does this mean that song is an abstract concept
not motivated by bodily experience? Yes, it does, because it is not universal bodily
experience that motivates it. Its motivation derives from the particular social-cul-
tural practice of the Ifaluk (p. 251).

Kvecses (2005) provides further examples of languages which show a cultural

basis for the time CM.
Metaphor across Time and Conceptual Space

Conclusion. However, in the last chapter of the book, he accepts the primacy of
embodied experience in conceptualization, at least for some metaphors:
My goal in this book is to offer a view of metaphor that can deal successfully
with the fact that some metaphors are potentially universal and the fact that some
metaphors vary cross-culturally and within culture (pp. 292293).

The immediate issue is the criteria for accepting a particular CM as universal. As

the review of the synchronic studies in the previous chapter of Japanese, Chinese,
Tunisian Arabic, Spanish, Russian, and Chagga have shown, emotion metaphor
employs significant cultural knowledge, and Kvecses own review of Gevaert
(2002) and Lutz (1988) came to the same conclusions.
Possible Solution. To answer the question concerning criteria for identifying
universal metaphor, Kvecses (2005) proposes, as one case, the angry person
is a pressurized container is potentially universal or near universal, be-
cause it is based in physiological experience and has been found cross-linguisti-
cally in a diverse set of languages (p. 64). Kvecses (2005) does acknowledge the
primary/complex distinction for the potentially universal group, but it is not
crucial in his view: In particular, these metaphors are simple or primary met-
aphors and/or complex metaphors that are based in universal human experi-
ences (p. 64); thus, in his view, both primary and complex metaphors can be
potentially universal.
Implications. If metaphors change over time as a result of cultural change, then
potentially universal is not a valid characterization of any CM. Because Kvecses
the angry person is a pressurized container CM is of the complex type, the
possibility is more likely (from Lakoff and Johnsons viewpoint) that cultural
knowledge influences the conceptualization. The larger issue is, therefore, the pos-
sibility of change over time in conceptualization via cultural models that provide
perspective on the experiential scene.

Geeraerts and Gevaert (2008)

In a later study by Geeraerts and Gevaert (2008), the authors specifically assert
that the claim of Kvecses (2005) for the potential universality of the the angry
person is a pressurized container CM is not supported by the results of their
study. The research investigated the semantic fields of words for heart and mood
in Old English, using Gevaerts (2002) data. The study method included an etymo-
logical study of compound words for heart (e.g., hatheort, hot-hearted) and
compounds for mood (e.g., tornmod, anger-mood/mind). Mood in OE literally
referred to the overall mental (cognitive and/or emotional) state of a person
(p. 323). In the first phase of the study, the etymological analysis of the compounds
Chapter 3. Metaphor across historical time

for heart appeared to lead to the conclusion, that in Old English, the heart is the
seat of feeling and thought (p. 323); that is, heart compounds refer to the heart as
the literal human mind. To corroborate this result, the compounds for mood were
compared to the compounds for heart, in order to confirm that the human heart is
conceptualized as the mind. The analysis revealed only 23 compounds for heart
and 78 for mood. The authors concluded that the asymmetry between the two
types of compounds indicates that the heart secondary in regard to
the literal denomination of mood, and an exclusive focus on imagery would sim-
ply distort the facts (p. 326). A third study of metaphorical anger revealed ten
etymological themes by which anger is expressed in Old English, with the result
that 65 of the expressions referred to literal anger and 56 referred to figurative
anger, though 14 of those were metonymic rather than metaphorical.
Study Conclusions. Geeraerts and Gevaert (2008) conclude from these three
studies that the etymological study of word senses in Old English revealed that
(1) mood was a literal expression for the full spectrum of mental life the rational
mind just as well as the feelings and the will (p. 339340); (2) heart compounds
comprise the same semantic range as the mood compounds; (3) yet, the heart im-
agery is a minor one compared to mood for referring to the mind. The authors
argue that the study method chosen revealed important details over an analysis of
the conceptual metaphorical mappings. Further, the authors state that certain con-
cepts, such as the etymological theme affliction, are conceptualized completely
independently of the pressurized container CM, indicating that the CM is not
universal (p. 342), contrary to Kvecses (2005) analysis discussed previously. For
English in particular, Geeraerts and Gevaert found that the pressurized con-
tainer CM is only weakly present, not a highly prevalent conceptualization, as
Kvecses claims (pp. 342343). Considering these findings, the researchers rec-
ommend etymological study of word meaning as a more effective means to dis-
cover the extent of both literal and metaphorical meanings of a word, compared to
a purely semantic analysis at the level of conceptual domains: an exclusive focus
on metaphorical conceptualization (a fortiori, on embodied metaphorical concep-
tualizations) is likely to distort the picture of the actual cultural models at work
(p. 343).
The study offers an important argument for investigating the influence of cul-
tural models on linguistic expressions. The existing theory or a particular type of
analysis can obscure the details of the relationship. This assertion is similar to
Deignans (2006) argument, discussed previously in this volume, that linguistic
theory should be applied to linguistic data, in order to investigate the validity of
the theory. From the point of view of the empirical study of language, these re-
searchers speak to an important methodological principle.
Metaphor across Time and Conceptual Space

Methodological Implications. We concur with the points made by Geeraerts

and Gevaert (2008); however, we further assert that linguistic data alone are not
sufficient to identify and analyze cultural models, as Lucy (1996) pointed out
(see Chapter 1). Conversely, linguistic theory alone is not sufficient to explain lan-
guage structure or cultural models (see discussion, this chapter). This is especially
true in diachronic studies because the extant cultural models are unfamiliar to the
contemporary researcher. Therefore, in the current study, linguistic data are sup-
plemented by non-linguistic data of historical cultural models. The non-linguistic
data precludes the need to appeal to theory alone to explain the study results and
adds information that is useful for analyzing the cultural models at work in the
linguistic data. We argue that this methodological solution has advantages over
appeals to theory or to historical linguistic data to explain research results. In ad-
dition, the Geeraerts and Gevaert (2008) study demonstrates that diachronic study
brings out details not available in synchronic data, and these features need to be
taken into account in synchronic research studies of conceptual metaphor.

Koivisto-Alanko and Tissari (2006)

Another recent study, which specifically investigated CM (Koivisto-Alanko &

Tissari, 2006), supported the conclusion by Gevaert (2002) concerning change in
semantic meaning over time, and added another important aspect change in CM
meaning over time. The researchers investigated the reason and emotion do-
mains by analyzing metaphoric expressions employing the words love, fear, wit,
reason, and mind. The words were searched in four corpora of English; two were
historical collections, the Corpus of Early English Sampler and the Helsinki Corpus
of English Texts; two modern-day collections included were the Freiburg-Brown
Corpus and the Freiburg-Lancaster-Oslo-Bergen (FLOB) Corpus. The purpose of
the study was to investigate the relationship between reason and emotion in CM
because, according to the authors, some researchers in previous synchronic stud-
ies argued that these concepts are similar, and others view them as divergent. The
researchers conducted a study to investigate the issue.
The total tokens collected for the five words were 2296 for love, 882 for fear,
1124 for reason, 1096 for mind, and 181 for wit. The analysis showed that Lakoff
and Johnsons (1980) ontological metaphor, in which an abstract concept (the tar-
get domain) is mapped by a physical entity (the source domain), was the basic CM
employed in the data. Specifically, for the five words studied, entity was the most
general source domain, and subdomains within it included container, instru-
ment/tool/weapon, obstacle, and valuable commodity. A second entity
was the human body; subdomains included the container for emotions and
reason. The CM identified in the analysis were love is a valuable commodity,
Chapter 3. Metaphor across historical time

the mind is a container, emotions are fluids in a container, several force

metaphors; in addition, the use of personification for abstract concepts was found
in human body metaphors, and quantification was employed in commodity
Study Results. The results showed that, historically, the container image
schema, body, and force/control were all used to map both reason and emo-
tion; conversely, body is used exclusively for the container of emotions.
Moreover, force/control is a continuum with reason on the control side and
emotions like love and fear on the force side; reason controls the extent to
which the emotions can surge (p. 209). Overall, in the CM for reason and emo-
tion the researchers found some marked differences (e.g., the force/control
continuum), and also some similarities (e.g., subdomains such as container and
force/control are shared).
The study also found that the metaphorical meanings of the CM change over
time. The authors discuss two types of change: (1) metaphor use to denote change
in the meaning of the expression, and (2) change in CM to denote cultural change.
The first type, change in the meaning of the expression, was found in the meta-
phors for wit. In the early modern English period, wit (the abstract target do-
main) was associated with mental activity is manipulation (the concrete
source domain), then later wit was associated by personification with a learned/
esteemed person, and finally wit became associated with the present-day mean-
ing of imaginative intelligence in the expression of speech and writing (Koivisto-
Alanko & Tissari, p. 210), indicating a change in the cognitive conceptualization of
wit over time.
Study Conclusions. The authors argue for two major points concerning the
results of the study. First, diachronic analysis of a concept can delineate variations
and changes in semantic meaning over time, as shown in the expressions employ-
ing wit. Second, CM change over time, and the changes reflect evolutionary
changes in cultural values. For example, the authors state that reason decreased
in its cultural value over time, indicated by its less frequent use in metaphor and
the restriction of its use to the philosophical text genre. Another example of cul-
tural change was found in the domain of fear. fear changed from possessing a
positive connotation in the Early Modern English period to a negative value in
present-day metaphoric expressions, indicating that emotions are evaluated dif-
ferently in different periods (p. 210), even within the same speech community. In
sum, high/low and positive/negative values are assigned to concepts by the speech
community using the concept, and these valuations constitute part of shared cul-
tural knowledge. This result delineates in specific terms the influence of a speech
community on cognitive concepts via the culturally-licensed cultural models in
use during the historical time period under study.
Metaphor across Time and Conceptual Space

Diachronic designs: Multiple factor studies

Several diachronic studies have found that there are multiple factors involved in
language change across time, similar to the synchronic studies discussed in
Chapter 2 (e.g., Barcelona and Soriano, 2004; Kvecses, 2005, 2009; MacArthur,
2005). Diachronic studies in this vein are Trim (2011) and Geeraerts, Gevaert, and
Speelman (2011).

Trim (2011)

Trim (2011) accepts many of Kvecses (2005) conclusions, including that the
angry person is a pressurized container CM is potentially universal, and also
accepts that culture is one of six parameters that affect the conceptualization of a
CM. Trim argues for an evolutionary model of diachronic change in conceptual-
ization and lists six parameters which can affect conceptualization over time: uni-
versal mechanisms, language, culture, conceptualization processes, semantic fields,
and salience (Trim, 2011, p. 2324). The six parameters discussed are reminiscent
of the multiple factors influencing CM discussed by Barcelona and Soriano (2004)
in Chapter 2, who recommended a multidisciplinary approach of cultural, neural,
psychological and linguistic accounts to investigate CM.4
Trim (2011) analyzes a variety of CM in English, including anger, love,
color, war, and shield for changes in the conceptualizations over time. The data
collected for the study are from literary sources, including Old and Middle English
texts, such as Chaucers The Canterbury Tales. Some data examples are provided
from non-English source texts that have been translated to English, such as a quo-
tation from Homers Iliad: He was full of melancholy; his innards swelled to dark
black. Trim uses this quote as evidence for cultural influence on CM by stating
that the color black is a reference to the Greek humoral theory (Trim, 2011, p. 71).
In addition, the quote is used to support Kvecses; proposed, potentially universal
pressurized container CM. Trim does not explain how the specific texts were
chosen for analysis.
Trim argues from his analysis that the shield concept has undergone a variety
of changes (205206) in both conceptual mapping and linguistic form, under the

4. There is some apparent overlap between Trims six parameters. For example, language and
culture overlap at various points, such as the social norms for communication in a speech com-
munity, and conceptualization processes and semantic fields overlap, at the level of CF theory, as
discussed in Chapter 1 for the constructs domain, base, ICM, and frame. Salience also overlaps
with culture; a conceptualization or a linguistic expression will be used more as its usefulness or
popularity increases its salience in a speech community (Boers, 1999); therefore, salience is the
result of cultural knowledge.
Chapter 3. Metaphor across historical time

influence of the six parameters. The universal parameter has preserved the protec-
tion = shield semantic field mapping, based on the visual perception of a shield
protecting the body in early warfare (203), and the other parameters have under-
gone a variety of changes over time. Trim concludes his study by stating that the
results of this study suggest that cultural features are usually a part of embodied
structures. The human mind needs to conceptualise within a cultural framework
which, as we suggested in the diachronic data here, is adapted to its age (p. 218).

Geeraerts, Gevaert, and Speelman (2011)

Another study that found effects for multiple factors in diachronic language
change, including culture, is Geeraerts, Gevaert, and Speelman (2011), in the lexi-
cal field of anger. This study employs the data collected by Gevaert in several stud-
ies to further study anger in English from 800 to 1500 A.D. in a multivariate study
design. The researchers investigated the research hypothesis of Diller (1994), who
suggested that, historically, a distinction between private and public expressions of
anger created the need for the development of the modern meaning of the lexeme
anger, a Scandinavian loanword. The prototypical meaning of anger before 1400
A.D. (trouble, affliction, vexation, sorrow) was gradually replaced by the increas-
ingly frequent use of the modern meaning (wrath, ire, hot displeasure) (Geeraerts,
et al., 2011, p. 110). The new meaning provided a means for differentiating public
and private anger as well as the social status of the person expressing the emotion.
For example, wrath traditionally referred to the violent expression of anger by a
person of high socioeconomic status; in contrast, the modern meaning of anger
referenced a non-violent (or at least, less violent) expression of anger by a person
of lower rank. Other studies cited also noted that, in historical texts, particularly in
Old English, only persons of power experienced anger. Based on these facts, the
researchers investigated the viability of Dillers hypothesis concerning the effects
of changes in culture on diachronic language change.
Data and Analysis. The database of English texts available for the study cov-
ered the historical period from 800 to 1500 A.D. The texts collected for the 1375 to
1425 A.D. period were selected, the time when the new meaning of anger first be-
gan to emerge in English. To investigate the research question, three types of anger
expressions were excluded from the texts, and those cases were eliminated from
the database: (1) expressions which did not evoke transitory emotional states;
(2) expressions in which the literal (non-metaphoric) expression was used as a
near-synonym; and (3), expressions in which it is stated that a person was not
angry. The instances were coded for two variables, text type and semantics. Text
type included two levels: religious vs. non-religious genre, and romance vs. non-
romance origin. The semantic variables included three levels: social status of the
Metaphor across Time and Conceptual Space

experiencer, private vs. non-private offense, and violent vs. non-violent expres-
sion. The researchers then conducted an exploratory bivariate analysis of the
effects of the three semantic factors, and a multivariate analysis of the combined
effects of the three semantic factors.
Study Results. The results of the two separate analyses indicate the following.
First, in the bivariate analysis, in the use of lexical items for anger expression from
1300 to 1500 A.D., there is a clear shift toward the use of anger in its modern
meaning of wrath, ire, hot displeasure. Second, all three of the semantic variables
have an effect on the use of the lexical item anger: the choice for anger is posi-
tively affected by the presence of non-high-ranking experiencers, private contexts,
non-violent reactions, non-religious texts, and non-romance origins of the texts
(Geeraerts, Gevaert, & Speelman, 2011, p. 123); in addition, the effects are statisti-
cally significant and support Dillers hypothesis. Second, the multivariate analysis
also shows support for Diller: ...the effect of a non-religious text, a private offence,
and a non-high experiencer is positive: these features favor the use of anger
(p. 124). It must be noted that this conclusion applies primarily to the non-reli-
gious texts of a non-romance origin.
Study Conclusions. The researchers conclude that the study supports the hy-
pothesis of Diller that persons of low social rank experience non-violent anger in
private contexts; in addition, the study found that text genre, particularly texts of
a non-religious, non-romance origin, affects the use of the non-romance loanword
anger. Though the study could not come to a definitive conclusion on this point,
two potential reasons are given for this effect of text type: for one, anger as a non-
romance loanword may significantly affect its distribution in the texts because a
word of non-romance origin would not appear in texts of romance origin, includ-
ing religious texts (pp. 123124). This conclusion is based on a consideration of
the text type variables.
Another possibility is that the modern meaning of anger arose largely in com-
municative contexts in which a commoner of low social status reacts in a non-
violent way to offenses that are considered private rather than public (p. 128); the
use of the word anger increases over time in concert with this particular meaning
(p. 121). This conclusion is based on a consideration of the semantic variables in
light of social values and cultural practices. The researchers recommend further
study, particularly a larger number of lexical items, because more evidence is
needed to determine if the cultural interpretation is indeed viable.
Implications for the current studies. Overall, the study by Geeraerts, Gevaert,
and Speelman (2011) is important to the current study for several reasons. First,
the study results support the effects of culture on diachronic change in the lexical
field of anger, the semantic field that includes the metaphoric expressions under
study in the current work. Second, the specific finding that the modern (post-1400
Chapter 3. Metaphor across historical time

A.D.) meaning of anger is concerned with private, non-violent expressions of emo-

tion by persons of non-high rank is supported by the data collected in the current
study of English between 1500 and 1990 A.D. In particular, the spleen metaphor
data discussed in results of the macro-study of historical metaphor (see Chapter 6)
and the micro-study of spleen metaphor (see Chapter 7) display the same semantic
features. If this analysis is correct, then the CM of anger described by Lakoff and
Kvecses (termed in this volume the blood metaphor) employs the expression of
anger traditionally ascribed to the public, violent expression of emotion by a per-
son of high rank, and the spleen metaphor expresses the private, non-violent ex-
pression of emotion by a person of lower rank. The implication is that the spleen
metaphor is not a non-prototypical case of Lakoff and Kvecses anger metaphor,
but a separate prototype.
Specifically, the Geeraerts, Gevaert, and Speelman (2011) study supports
the possibility that the two types of metaphor may be motivated by different
experiential scenes, as Chapter 2 postulated, and if so, the spleen metaphor
and the blood metaphor are interpreted by different perspectives (i.e., cultural
models), producing two different conceptual domains, mappings, and CM. This
result would further indicate that the spleen metaphor constitutes a second pro-
totype of anger. These issues will be discussed again in the final chapter of this
Universal aspects of embodiment have been found across many languages, but
cultural aspects also are found in the same data, as shown in the literature reviews
of synchronic and diachronic studies of CM. We argue that the CM studied in
previous synchronic research was influenced by diachronic cultural models, if the
principle of non-autonomous knowledge is to be taken seriously. When culture as
a variable is not a goal or object of a particular study, then cultural models will be
overlooked5 and their influence on conceptualization will be lost.


Cultural knowledge has been viewed as an important factor in historical studies of

emotion expressions, including conceptual metaphor. The results of the diachron-
ic studies of CM, including Gevaert (2002), Geeraerts and Gevaert (2008),
Geeraerts, Gevaert, and Speelman (2011), Koivisto-Alanko and Tissari (2006),
and Trim (2011) indicate that the use of a cognitive concept over time may affect

5. See Domaradzki (2011) for discussion of the effects of universalist and relativist viewpoints
on research methodology.
Metaphor across Time and Conceptual Space

its meaning, and that evolving changes in cultural values may affect the conceptu-
alization of a CM. The exploratory study by Geeraerts and Grondelaers (1995)
discussed the possibility of these relationships, and the later corpus-based studies
by Gevaert (2002), Geeraerts and Gevaert (2008), Geeraerts, Gevaert, and
Speelman (2011), Kvecses (2005), and Koivisto-Alanko and Tissari (2006), have
provided details which support the relationship between embodied experience
and cultural models.

The research gap

However, as mentioned previously, CF research has not often investigated change

over time in linguistic metaphors of anger, which is the theoretical foundation
of embodied metaphor in CM theory. The Geeraerts, Gevaert, and Speelman
(2011) study indicates that the prototypical meaning of anger has changed in the
past 500 years; in addition, the study implies that the CM of anger is motivated
by a particular experiential scene that includes cultural variables, such as the so-
cial rank of the experiencer and the public or private nature of the offense that
causes the expression of emotion. This conclusion supports the analyses in
Chapter 2 of this volume concerning the divergent features of the Anger Proto-
type Scenario for the spleen metaphors. In fact, it appears that the comparison of
the blood and spleen metaphors indicates that each is motivated by a different
experiential scene and therefore linguistic metaphors for each type are instanti-
ated by different cultural models, domains, mappings, and conceptual metaphors.
In addition, this conclusion indicates that blood and spleen metaphors constitute
separate prototypes of anger.6 The existence of multiple prototypes for anger
also provides evidence to support the existence of a domain matrix of emotion,
as discussed in Chapter 1.
If these results are caused by changes in cultural beliefs and values over time,
then that would indicate that cultural knowledge is an important factor in concep-
tualization generally and in CM specifically. Cultural knowledge is employed to
select culturally-licensed dimensions of a CM for encoding in the semantic mean-
ing of metaphoric expressions, in the same way that cultural knowledge selects
culturally-licensed dimensions of the experiential scene and encodes the scene in
the syntax of linguistic expressions.

6. The existence of multiple prototypes for an emotion concept has been found in other stud-
ies of emotion concepts. Kvecses found three separate prototypes of happiness (1991; see also
2010b, p. 113) and provides several justifications for that conclusion, including the assertion that
other emotion concepts consist of more than one prototype (1991, p. 44).
Chapter 3. Metaphor across historical time

Theoretical implications

An important implication of the above discussion is that, over time through re-
peated use of perspective to encode the scene, cultural models may independently
change the nature of the conceptualization. This specific idea is not currently ac-
cepted in CMT. For example, Kvecses (2005; 2009) states that cultural models de-
velop out of and are determined by a particular conceptual metaphor; therefore, a
model may only temporarily override a conceptualization of embodied experience.
The implication is that the cultural model cannot permanently change the embod-
ied concept, and this limitation provides the theoretical grounds for a potentially
universal metaphor. However, recent researchers discussed above have found some
evidence that change in a conceptualization can occur via cultural models, and the
current studies investigated this question, as well. Crucially, we will argue that the
influence of cultural models (or override, to use Kvecses term) is not temporary,
as Kvecses pressure of coherence model suggests, but is permanent as long as a
particular cultural model is active in a particular speech community.
These studies as a whole call into question the universal, pre-cultural charac-
teristics of conceptualization theory (discussed in Chapter 1), as evidenced by the
changing frequency of use of the conceptualization and in its content (construal).
If CMs change as a result of use and changes in culture, when, if ever, would a CM
display universal characteristics? All CMs, constantly situated in a cultural milieu,
may consistently obscure or selectively ignore some or all of their embodied expe-
rience; some CMs may do this relatively more or less than others, but universal
cultural experience must affect all CMs in some way, if it affects any of them. In
sum, the studies reviewed above indicate the importance of studying CMs dia-
chronically, as well as point out some important issues that are difficult to account
for within current conceptual metaphor theory. In addition, several studies, in-
cluding Geeraerts and Grondelaers (1995) and Gevaert (2002), cultural knowl-
edge of the human body specifically from the field of medicine (especially the Four
Humors model for English) contributed important aspects of the conceptualiza-
tion of the body and emotion.

Methodological implications: The role of frequency statistics

The studies also provide research designs that are instructive for the study of con-
ceptualization and cultural models. Geeraerts and Grondelaers (1995) use of non-
linguistic data for metaphor analysis is useful for the study of cultural models and
their current effects on language. Gevaerts (2002) frequency analysis is interesting
for its implications for research methods, and Geeraerts, Gevaert, and Speelmans
(2011) use of lexical field analysis is useful to identify semantic and cultural
Metaphor across Time and Conceptual Space

variables. In CM studies, frequency analysis has not been employed widely, pos-
sibly as a result of the widespread use of the introspection method for collecting
and analyzing language data as well as the general view that frequency does not
necessarily indicate prototypicality (see Kvecses, 2008; 2011). Yet, accepting this
view as a valid methodological caution does not preclude the potential use of fre-
quency data in CM research changes in frequency may indicate a change in the
currently-licensed cultural model which in turn may signal a change in the proto-
typicality of a lexeme, linguistic expression, and/or a CM. Several of the studies
discussed above employ frequency analysis in scientifically-valid ways to investi-
gate prototypicality and found several such changes.

Motivations for the current studies

On the basis of this review of the research to date, a full investigation of the influ-
ence of the Four Humors cultural model on linguistic expressions which instanti-
ate the CM of anger has not been conducted. Geeraerts and Grondelaers (1995)
was exploratory and synchronic-historical, not diachronic, and Trim (2011) em-
ployed preselected data removed from the historical texts, possibly obscuring the
relationship between the linguistic expression and the cultural and social context
in which the expression occurs. Gevaerts (2002) diachronic study looked at
the conceptual domain of anger via individual lexical items, not the CM of
anger in metaphoric expressions. Geeraerts and Gevaert (2008) investigated the
semantic fields of words for heart and mood in Old English for the CM of anger,
rather than the full linguistic expressions in context, and Geeraerts, Gevaert, and
Speelman (2011) studied change over diachronic time in the lexical field of anger
rather than the specific features of the CM of anger. Finally, Koivisto-Alanko and
Tissari (2006) investigated emotion CM, not the CM of anger. Therefore, no
previous diachronic study of the CM of anger has investigated cultural change in
a corpus-based research design of language data in its full linguistic context, track-
ing the frequency of use of the CM of anger and analyzing its full semantic context
to indicate changes in the conceptualization over time. Considering the impor-
tance of the CM of anger both theoretically and historically to CMT, this gap in
the research needs to be addressed.
Given the goals of CF research to understand the human mind and the advan-
tages that diachronic research brings to understanding the influence of culture on
conceptual metaphor, adding time as a variable to the study of CM enhances the
researchers ability to understand variation in cognitive conceptualization. For ex-
ample, synchronic studies show that that the two types of deictic orientation exist
in the worlds languages, but not how they developed. Diachronic study could de-
lineate changes in experience and cultural values over time which led to the
Chapter 3. Metaphor across historical time

development of variations in present-day conceptual metaphors. The current study

addresses this methodological issue by employing a diachronic research design
covering almost 500 years of the English era. By this procedure, changes in con-
ceptual metaphors that result from changes in cultural knowledge were delineated
more clearly. Chapter 4 describes the research design for the ancillary study of
non-linguistic data and the main study of linguistic expressions of anger.
part ii

A macro-study of human emotion in cultural

context, A.D. 15001990
chapter 4

Research questions and methodology


This chapter will discuss the research design and method for the five-century study
of historical metaphors of anger. The chapter is divided into four parts: research
questions, the methodology for the ancillary study of non-linguistic data, the
methodology for the main study of historical linguistic expressions of anger, and
the chapter summary. Data collection and analysis, along with the results of the
ancillary study, are described in Chapter 5. The results of the main study of dia-
chronic metaphors of anger are discussed in Chapter 6.

Research questions

The research questions were developed from the results of the literature review of
synchronic and diachronic conceptual metaphor studies, discussed in Chapters 2
and 3.
1. What was the conceptual relationship between the blood metaphor and the
spleen metaphor during the historical period? Are they located in the same
CM, different CM, or is the relationship characterized in some other way?
2. What motivated the conceptualization in each type of metaphor? Is it bodily
experience, cultural knowledge, a combination of these, or some other
3. Did changes in cultural models in turn change the conceptualization of anger
over time?
4. Did scientific knowledge (and advancement in that knowledge) influence the
cognitive conceptualization of anger?
The next section provides an overview of the methodology of the ancillary study
of historical non-linguistic data, conducted to collect information on the Four Hu-
mors cultural model.1 This data was used to aid the accurate interpretation of the
linguistic data collected for the main study of diachronic metaphors of anger.
After the section detailing the ancillary study, the main study is described.

1. See Chapter 1 for discussion of non-linguistic data and its importance to the current study.
Metaphor across Time and Conceptual Space

The ancillary study of historical non-linguistic data

Data collection

The methodology for the collection of the non-linguistic data was patterned after
the method employed in Geeraerts and Grondelaers (1995).2 Examples of the use
of the Four Humors cultural model in scientific texts, literature, art, and music
were collected to investigate the knowledge the speech community of the time had
of the cultural model and to use this non-linguistic data to interpret the semantic
meaning of the metaphorical expressions collected from the corpora. In the main
study, the CADS method and the non-linguistic data were used to analyze and in-
terpret the diachronic conceptual metaphor samples collected from text corpora.
Four types of information were collected for the historical study, following the
discussion of metaphor research methodology in the preceding section. Recall
from Chapter 1 that the study of culture in language may be conducted using lan-
guage data, non-linguistic background data, or both. As discussed in that chapter,
both types were selected for the current studies because metaphoric expressions
occur infrequently, resulting in a low number of samples that may not provide a
detailed picture of the effect of culture on language. The historical background
data were added to increase the breadth, depth, and accuracy of the analysis of the
main study of conceptual metaphor of anger.
The four types of non-linguistic data collected for the ancillary study were
(1) historical sources on the Four Humors scientific theory (16th and 17th centu-
ries); (2) Four Humors cultural practices (16th to 20th centuries); (3) historical
scientific advances in human physiology (16th to 20th centuries); and (4) linguis-
tic samples from corpus data samples that explicitly mention the Four Humors
model. Each of these types will be discussed in turn below. After the descriptions,
the collected data are discussed in one century time frames to develop the com-
posite model and view changes in the Four Humors model over time.

The four types of data

1. Historical Sources on the Four Humors Model of Human Health

The information concerning the Four Humors focuses on the historical period
between 1500 and 1700 A.D., when the model was an important influence in
medical practice and culture in Western society. The scientific and cultural
importance of the theory at the time was on par with the models standing
during its first golden age in classical Greece, the original source of the theory.

2. See Chapter 3 for discussion of this study.

Chapter 4. Research questions and methodology

Nutton (1995) states that for the Greeks, the humoral system was capable of
almost infinite variation, unfalsifiable on its own terms, and often correspond-
ing to the facts of observation (p. 25). The same can be said for the expanded
model developed by Renaissance thinkers, who also increased the models ex-
planatory power to include all aspects of the known universe in the historical
time period.
The Selection Method. The method for selecting the historical source texts for
the ancillary study of non-linguistic data included the following procedural
steps. First, experts who described important elements of the model were the
subject of a library search of secondary historical sources, to cull the funda-
mental principles which guided the theory and its application by lay medical
consumers during the two-century (i.e., 15001700 A.D.) period. Annotated
bibliographies of historical Four Humors texts by Draper (1945) and Babb
(1951) served as a starting point for the library search. Candidate texts were
obtained either in paper form from the Newberry Library in Chicago or in an
electronic facsimile version from Early English Books Online (EEBO). Docu-
ments selected for the study were chosen for their detailed descriptions of the
Four Humors and their popularity among a wide variety of readers, in order to
select texts that had an influence on lay practitioners of the model during the
historical period. In sum, the documents were chosen for their ability to influ-
ence the knowledge and values of a broad section of English-speaking society.
Criteria for Selecting Four Humors Texts. The specific criteria for selecting cul-
turally-significant documents included the following: each text (1) was writ-
ten in or translated to English for a general (non-professional) audience; (2)
was reprinted at least twice, for a minimum of three printings; and (3) in-
cluded discussion of the basic tenets of the Four Humors model. The first cri-
terion was further delineated by the following sub-criterion: the text did not
use any Latin words or phrases, a sign that the text was written for lay people,
not experts. The third criterion also had a sub-criterion: specific information
was provided in the text on the following fundamental principles of the model,
identified in historical research the four humors, the four qualities, the four
temperaments, and the four principle organs.
Results of the Selection Procedure. In all, 31 texts which were published during
the two-century period were analyzed as candidates for the ancillary study;
some were eliminated from consideration for the ancillary study because they
did not fit the selection criteria. First, four texts were categorized as profes-
sional books, and eight others had fewer than three editions; these were elimi-
nated following the selection criteria identified previously. However, one book
which did not meet the second criterion (i.e., fewer than three printings) was
placed on the list of texts used in the study, due to its seminal and detailed
Metaphor across Time and Conceptual Space

discussion of the Four Humors model.3 Finally, two texts which met the first
two criteria were eliminated for not meeting the third criterion lack of agree-
ment with the other sources on key points in the model or for not discussing
one or more of the fundamental principles of the model. For example, one text
did not accept the principle of the operation of the four qualities (wet/dry, hot/
cold) on the four humors. The selection procedure identified a total of 18 his-
torical source texts for the study. Data collected from the selected sources
include a brief annotated bibliography (See the References Section) and a table
of the selected texts with summary data on the Four Humors basic principles
discussed in each text. Finally, a detailed description of the Four Humors
model, called the composite model, was developed from the information found
in the selected sources.4
The summary data (See Table 1 at the end of this chapter) showed that all were
original works in English or translations to English. The oldest text was pub-
lished in 1542, and the latest was a book reprint, published in 1698. All of the
documents were authored and/or published in England, but other countries in
Europe contributed authors, including France, Spain, Denmark, and Italy. The
authors included physicians writing medical treatment texts for laymen, aca-
demics discussing different aspects of the known universe and the practical
effects on human life and health (e.g., the influence of the stars and planets on
childbearing), and religious authors writing about the relationship between
the human body and spiritual life. Moreover, all of the books discuss the Four
Humors in enough detail to delineate the primary features of the model. There
are some small disagreements among the writers on certain details of the
model, such as recommended treatments for a specific illness, yet all agree on
the fundamental tenets. Overall, the texts selected represent a composite view
of the Four Humors model as it was generally constituted and practiced by
physicians, clergy, and lay practitioners during the English Renaissance.
2. Historical Data on Four Humors Cultural Practices
The second type of historical data collected was information on specific cul-
tural practices in English-speaking society that were initiated by shared cul-
tural knowledge of the Four Humors scientific theory. This type of data showed
that the Four Humors had penetrated society to the point of becoming a

3. This text is a 1582 edition of Bartholomew de Glanvilles De proprietatibus rerum, origi-

nally written in Latin in 1360. Though the original work is not a product of the historical period
under study, the text was included due to its position of authority for the Renaissance Four Hu-
mors authors, who often used de Glanvilles work as source material for their own treatises.
4. The table of summary data (Table 1) is located at the end of this chapter; the composite mod-
el is discussed in Chapter 5. Both the table of summary data and the composite model were em-
ployed in the analysis of the data in the main study of historical metaphor.
Chapter 4. Research questions and methodology

culturally-licensed social practice among lay medical consumers. Acceptance

of the model by people who are not medical experts would suggest that the
theory influenced not only expert theories but also the values, activities, and
language of the general population. Thus, historical data provided evidence of
the influence of the Four Humors medical model on society, and such influ-
ence included the use of language about physical and emotional health in
metaphoric expressions of anger and other emotions.
The procedure for collecting the cultural data was as follows. The Newberry
Library in Chicago was visited on two separate occasions to consult the li-
brarys collection on the Renaissance and Modern English periods (i.e., 16th
through 20th centuries). The holdings include medical books employing the
Four Humors model; scientific treatises which applied the model to a specific,
contemporary issue or problem; theological sermons; personal diaries and
correspondence; fictional novels, poems and plays; and, artwork and music.
Specific works from each of these cultural sources were analyzed for descrip-
tions of Four Humors cultural practices in the everyday life experience of lay
people. The information collected was employed to analyze the knowledge
that lay people had of the model and how that knowledge was used in every-
day life.
3. Historical Data on Scientific Advances in Human Physiology
The third type of data collected for the ancillary study concerned the changes
in knowledge of the human body and its physiological processes over the five-
century period. This data type was important because the CM of anger
contains many concepts about the structure of the human body and related
biological processes. Secondary library research was conducted to collect data
on important scientific advances during the 490 year period under study. The
scientific advances were selected by conducting research in the field of the his-
tory of medicine to determine which advances were considered by medical
historians to be the most influential on scientific knowledge of the period; one
major advance related to human physiology for each of the five centuries was
chosen, for a total of five. The five major advances were placed on a timeline
and compared to metaphor data samples from the same period to investigate
if there were changes in language meaning or use that corresponded with the
timing of each scientific advance. The purpose of the procedure was to view
changes in language that result from changes in lay understanding of scientific
knowledge of the body, indicating a possible change in cultural models that
would affect metaphorical expressions of the body.
4. Non-metaphorical Data Samples Referencing the Four Humors Model
Finally, the fourth type of data collected were non-metaphoric linguistic ex-
pressions, found in the compiled corpora used for the main study of diachronic
Metaphor across Time and Conceptual Space

metaphor. The data were collected while conducting the keyword searches of
the corpora for the main CM of anger study (see Chapter 6 for details). The
non-metaphorical language samples were collected from the pool of linguistic
metaphor samples; however, the meaning of the word or phrase in the specific
situational context is not a source-target mapping of the human body and
emotion. Instead, the reference is to the actual practice of the Four Humors
model during the historical period. These non-metaphorical language samples
were used to provide further evidence of the cultural values and practices of
the model during the historical period under study.

Data analysis

The four types of data collected for the ancillary study were analyzed for two main
purposes. First, the 18 historical sources on the Four Humors model were used to
create a detailed description the composite model of the Renaissance Four
Humors model. The composite model was compared to the linguistic expressions
collected in the main study of diachronic metaphors of anger to view changes in
the conceptualization of the emotion over time. Second, the other three data types
(i.e., historical cultural practices, scientific advances, and non-metaphorical
linguistic samples) were compared to the linguistic expressions collected in the
main study to investigate the effect of cultural change on conceptualization; as
cultural practices and scientific knowledge changed over time, the effects of
cultural change on the instantiation of the CM of anger in linguistic metaphor
were delineated.5
The results of the ancillary study were used to aid the interpretation of the data
collected for the main study of diachronic metaphors of anger. The following sec-
tion describes the methodology for the main study.

The main study of diachronic linguistic expressions of anger


Two compiled text corpora of historical English texts were selected for the
study; the corpora included the Penn-Helsinki Parsed Corpus of Early Modern
English (PPCEME) and A Representative Corpus of Historical English Registers
(ARCHER). The PPCEME included 1.8 million words, a compilation of texts from

5. The results of the analysis of the ancillary study are presented in Chapter 5.
Chapter 4. Research questions and methodology

a wide variety of genres, dated between 1500 and 1720 A.D.; it is distributed on
CD-ROM by the University of Pennsylvania. The ARCHER corpus included 1.9
million words, compiling texts between 1650 and 1990.6 Details on the types of
texts and the word counts are available in Appendix A for the Penn-Helsinki cor-
pus and in Appendix B for the ARCHER corpus.
In addition to the two corpora described above, two other computerized text
collections were employed to provide additional keyword samples. The Modern
English Collection of the University of Virginia Electronic Text Center was searched
for the 15001899 period, and the Making of America Collection of 19th century
British and American magazines at Cornell University was searched for the 1800
1899 period. The purpose of the additional examples was to explicate in more de-
tail particular features found in the Penn-Helsinki and ARCHER corpora.7


Based on the research review in Chapters 2 and 3, we propose that the specific
nature of the non-prototypical cases in Lakoff and Kvecses (1987) data needs to
be investigated because of the possibility that there were two different prototypical
conceptualizations for anger in English during the historical period under study.
To investigate the question, the current study collected samples of two metaphoric
expressions in natural language data from compiled, historical corpora. The ex-
pressions chosen are given in the examples below.
A. His blood boiled with rage.
B. He vented his spleen on the unfortunate man.
Sample A (called in this study the blood metaphor) instantiates the CM of anger,
as analyzed by Lakoff and Kvecses (1987). Sample B represents the controlled
response over time group of expressions, discussed in Chapter 2 (called the
spleen metaphor). The underlined portions indicate the basic metaphoric expres-
sions. Lakoff and Kvecses propose that the spleen metaphor is an extension of the
CM of anger; however, these two cases have been shown in my own analysis to
have characteristics which are significantly different in comparison. In addition,
there is also the possibility of a domain matrix, a group of interrelated conceptual
domains, in which the blood and spleen metaphors are separate dimensions; see

6. At the time of the data collection, ARCHER was housed in a computer database located at
Northern Arizona University.
7. These samples are not numbered and are discussed in this volume only as needed; see
Chapter 6.
Metaphor across Time and Conceptual Space

Croft, 2003/1993; Langacker, 1987. The data collected and analyzed for the current
study explored these questions.

Implications for data identification and collection

As mentioned previously, the metaphoric expression for spleen shown above (the
B sample) presupposes the target domain of anger, rather than referring to the
target explicitly through the use of a lexical item. An implied target domain is
common but problematic in linguistic metaphor. In contrast, an explicit target is a
relatively easy search strategy in compiled corpora because all metaphoric expres-
sions contain lexical items pertaining to the source domain (Stefanowitsch, 2006b),
and identifying the words that are used in a particular source domain generally
results in a shorter list of lexical items than the list for the target domain. When the
target domain is not represented by a lexical item (the implied type), there are two
problematic issues for research methodology: (1) the search procedure requires an
exhaustive list of source domain words, in studies in which the objective is to iden-
tify every instance of a metaphor that is relevant to the target domain; and, (2) the
researcher must rely on his own competence in the language under study to deter-
mine the target domain from the content and meaning of the metaphor, a problem
that Enfield (2000) and Lucy (1996) state must be mediated in research studies of
language and culture through the use of non-linguistic data (see Chapter 1). This
second issue is particularly critical in studies of historical metaphor because the
researchers own competence in the language is not sufficient to determine the
target domain for a historical speech community. These two weaknesses were ad-
dressed in the design of the current study.
For the reasons given above, Stefanowitsch (2006a) suggests that research de-
signs that employ compiled corpora should search for instances of a metaphor in
which both the source and target domains are lexicalized; he calls this type of
metaphor a metaphorical pattern. This class of metaphorical expression is more
straightforward to analyze than those with implied targets, but generally, there are
fewer instances of the metaphorical pattern type than the implied target type, and
fewer cases may affect the results of the analysis. Since metaphor is a relatively rare
phenomenon in texts (and text corpora), reducing the number of samples col-
lected may also affect the ability to generalize the study results. Understanding
these issues, a methodological approach that deals with the problems of implied
targets yet does not reduce significantly the number of data samples was needed
for the current study. The method chosen was the use of a limited number of key-
words to search for metaphor samples in the corpora.
Chapter 4. Research questions and methodology

Selecting the search keywords

To address the issue of implied targets, the list of source domain lexical items was
limited to four keywords vent-, spleen, blood, and boil-.8 There are several advan-
tages to collecting language samples using a small number of selected keywords.
First, these are the lexical items that comprise the A and B samples of the meta-
phoric expressions of anger (see previous discussion) chosen for the current study,
ensuring that that the implied target domain of anger is included in the collected
samples. Second, the number of lexical items to be searched in the corpora is
small, simplifying the search procedure. It is true that the items selected will not
result in an exhaustive collection of metaphoric expressions for the domain
(as Stefanowitsch, 2006a recommended); however, the purpose of the current study
is to investigate change in the blood and spleen metaphors, not to compile a com-
plete accounting of the lexical items employed to infer the target domain of anger.

The role of context

In addition, some of the implied targets were recovered during the data analysis
phase because the text corpora chosen include the full context within which a
linguistic expression occurs. In some samples, the target domain was located in a
sentence adjacent to the sentence which contains the metaphorical expression. By
using the keyword search procedure in full-context corpora, the negative effects of
implied targets were largely eliminated, data samples were not reduced signifi-
cantly, and the total number of samples increased via contextual information.

Limitations of the keyword search procedure

One consequence of the above keyword search procedure is that several cases in-
volving metaphorical use of the keywords ultimately were not analyzed because
the implied target was not an emotion. This decision was made in order to focus
on the uses of the keywords which related directly to the CM of anger. Koivisto-
Alanko and Tissari (2006), discussed in Chapter 3, used a similar procedure; the
researchers analyzed only the cases of their selected keywords (i.e., fear, love, mind,
reason, and wit) which related directly to the target domains reason and emo-
tion. This procedure was employed in order to investigate the researchers spe-
cific research questions; as a result, This means we have deliberately left out some
central metaphors that did not fit within the scope of this study (p. 194). Since the

8. The two words with the dash (-) denote verbs which have variable inflectional suffixes in
linguistic metaphors; the other two words are mass nouns which typically are not inflected in
linguistic metaphors of anger.
Metaphor across Time and Conceptual Space

research questions for the current study focused on the blood and spleen meta-
phors, a procedure similar to Koivisto-Alanko and Tissari (2006) was employed:
only the keyword cases for vent-, spleen, blood, and boil-, which clearly instantiated
anger, were analyzed. It was expected that the keyword cases collected would be
useful ones for investigating the specified research questions.

Data collection

The data samples for the four selected keywords were collected from the PPCEME
and ARCHER corpora, preserved in their original discourse context, during the
time period between 1500 AD and 1990 A.D. The use of the two compiled corpora
allowed for the analysis of changes in the CM of anger for the entire modern
English period. In addition, to analyze the data in greater detail, the non-linguistic
background data on the Four Humors medical model, cultural practices associ-
ated with the model, and historical scientific advances that may have brought
about changes in cultural beliefs regarding the Four Humors were used to aid the
analysis of the metaphor data and the CM. Finally, in certain cases, the two addi-
tional digital corpora (i.e., the University of Virginias Electronic Text Centers
Modern English Collection and Cornell Universitys Making of America Collection)
were searched to provide additional examples of particular features of the meta-
phoric expressions.

Data analysis

Four steps were involved in the data analysis procedure: (1) classification of the
data; (2) identifying instances of metaphorical expressions; (3) calculating fre-
quency of use statistics for the metaphors; and, (4) analysis of both the metaphor-
ical and non-metaphorical cases to investigate the research questions. Each of
these steps will be discussed in detail below.
1. Classification of the Data
For the first step, the keyword instances were classified according to the date
of publication and placed in chronological order by corpus (i.e., either the
Penn-Helsinki or ARCHER). This information was available in the back-
ground data on each text when each corpus was compiled (see the Appendices
for summary information for the two corpora).
2. Identifying Instances of Metaphorical Expressions
Second, the keyword instances were read in their original context in order to
determine which cases constituted metaphorical expressions. Recall that, in
this study, a linguistic expression is considered an instance of a metaphoric
Chapter 4. Research questions and methodology

expression if a source domain concept (e.g., container) is mapped onto a

target domain of emotion (e.g., anger). The target domain could be referred
to explicitly by a word or implied by the context; applying this definition to the
collected cases, both His anger boiled up and His blood boiled were considered
instances of metaphoric expressions.9
There were two situations in which the identification of metaphoric ex-
pressions became difficult: cases in which the target domain was not referred
to explicitly in a word or phrase (i.e., an implied target; see previous discussion
in this chapter), and cases in which the interpretation of discourse context
determines the correct reading. The correct reading was difficult in some cases
because determining what is an emotion and what is not is sometimes not
clear or the boundaries between different emotions overlap (Koivisto-Alanko
& Tissari, 2006). Four samples from the collected data will be discussed to il-
lustrate these two problems in the metaphor identification process employed
in the study. First, a clear case of a metaphoric expression of emotion, from a
news report in the Penn-Helsinki corpus, is shown below.
A. And indeed mens spirits were so sharpened upon it, that we all looked on
it as a very great happiness that the people did not vent their fury upon the
papists about the town. (1724)
Case A is clearly a metaphorical expression; the structure employs the same
one described in the model discussed previously (i.e., He vented his rage), and
the human body source domain (container) releases the emotion target do-
main (anger). This is a case of a metaphorical pattern, as described by
Stefanowitsch (2006a).
In contrast, Case B is a clear case of a non-metaphorical meaning from a 16th
century medical text:
B. ...and this Arteir carieth blood from the Hart to the Lungs, the which
Blood is vaporous, that is tried and left of the Harte, and is brought by this
Artery to the Lunges, to geue hym nutriment. (1548)
The structure of the sample does not employ the model of a metaphorical ex-
pression, and more significantly the source domain (blood) is not mapped to
a target domain of emotion, but is discussed in terms of other entities the
physical location of blood in the human arteries and its role to nourish the
lungs. Based on the analysis, the A and B cases are classified as metaphorical
and non-metaphorical expressions, respectively.

9. Cases which instantiated emotions other than anger were also used to investigate the con-
ceptual relationships between emotions. See Chapter 6.
Metaphor across Time and Conceptual Space

As mentioned previously, some cases were difficult to classify because the tar-
get domain was implied, or the context of the expression was difficult to inter-
pret accurately. Case C shows an implied target.
C. He was forced to retire to vent his groans, where he fell down on a carpet,
and lay struggling a long time, and only breathing now and then Oh
Imoinda! (1688)
Here, the structure of the model is present, and the container source domain
is employed, but the emotion target domain is implied rather than expressed
explicitly. The context was then analyzed to provide material for accurate in-
terpretation. The man exhibits behaviors (i.e., reclining on a carpet, groaning,
repeating a womans name), which point to despair over a lost love, so the case
was classified as a metaphorical expression of sadness.
Finally, some cases were difficult to classify because the context of the expres-
sion was unclear. Case D below is an example.
D. ...but of all Creatures I hold that Wife a most vnmatched treasure, That can
vnto her fortunes fixe her pleasure, And not vnto her Blood, this is like
wedlocke, The feast of marriage is not Lust but Loue, And care of the es-
tate, when I please Blood, Meerely I sing, ... (1630)
The model structure is present (I please blood), and the source domain is the
action of serving; however, the target domain is implied. In the context of
marriage and love, blood in this case appears to refer to sexual desire personi-
fied as the one the speaker serves. The key analysis question then becomes, is
sexual desire an emotion, and/or does the conceptualization overlap with
emotion categories (Koivisto-Alanko & Tissari, 2006)? In this case, the answer
was Yes because sexual desire overlaps with love, and the word love appears
in the sample, corroborating that the emotion was the intended target domain
(though the speaker is contrasting the differences between love and sexual
desire in the sample). In addition, CM researchers have noted the use of per-
sonification in linguistic metaphor (e.g., Lakoff & Johnson, 1980; Lakoff &
Turner, 1989; Koivisto-Alanko & Tissari, 2006). Overall, based on the evi-
dence, Case D was classified as a metaphorical expression. To summarize, the
procedure described above was employed to identify the cases which were
metaphorical expressions of emotion; the remainder were classified as non-
metaphorical cases.10
3. Calculating Frequency of Use Statistics

10. A few cases were classified as metaphorical but did not instantiate emotion, and these were
eliminated from the analysis; see the previous section on Data for further discussion of this
Chapter 4. Research questions and methodology

In the third step of the data analysis, some statistics on frequency of use of the
keywords were generated. Frequency of use was identified in Chapter 2 as an
important factor for the study of cultural models; Gevaerts (2002) study of the
conceptual domain of anger also showed that changes in the frequency of use
of a lexical item may indicate a change in the cultural value of the concept as-
sociated with the word. To study this factor in the current study, the frequency
of use of the four keywords in metaphorical expressions was counted and
tracked over time. Raw frequency counts were normalized to the rate of occur-
rence per one million words of running text, in order to allow for comparisons
between historical periods. The formula, from Biber (2006), is as follows:
raw frequency x 1,000,000
total words in corpus
The Normalized Frequency Rate (NFR) for each historical period indicated
the frequency of occurrence of the metaphorical use of a keyword during the
period, which is in turn a measure of the relative importance of the expression
in the language at that time.
In order to calculate the NFR accurately for the two corpora, some meta-
phorical samples were eliminated from the dataset. The reason is that the two
corpora overlap between 1650 and 1720; including all of the metaphoric sam-
ples from both corpora would skew the NFR calculations for that historical
period. To resolve the overlap, the Penn-Helsinki samples from 17001720
(three total cases) were eliminated, and the ARCHER samples between 1650
and 1699 were also eliminated (seven total cases). The procedure created a
defined period for the collected metaphor data from each corpus: the Penn-
Helsinki data covered 1500 to 1699 A.D., and the ARCHER data covered 1700
to 1990 A.D. The eliminated samples were not included in the calculation of
the NFR; however, they were retained for the purpose of analyzing the struc-
ture of the conceptualization of anger.
For the two additional corpora, the Modern English Collection of the Uni-
versity of Virginia Electronic Text Center and the Cornell University Making
of America Collection, frequency statistics could not be generated because
these two collections did not track the total number of words in the collection
or in individual documents. These corpora were used only to aid the analysis
of particular features of samples found in the Penn-Helsinki and ARCHER
4. Analysis of the Metaphorical and Non-metaphorical Samples
Finally, in the fourth step of the analysis, two types of analysis techniques
one quantitative and one qualitative, following the CADS analysis system
were employed for the samples of metaphoric expressions.
Metaphor across Time and Conceptual Space

Quantitative analysis. The raw frequency counts and the NFR were analyzed to
study the changes in the frequency of use of the metaphoric expressions; combin-
ing this data with the specific date that a keyword began to be used and/or fell out
of use showed the arc of the historical use of a keyword in a metaphoric expres-
sion. The analysis, presented in 50-year increments (or cells) from 1500 to 1990
A.D., delineated correlations between frequency of use and the rise and fall in
popularity of the Four Humors model in English-speaking society.
Qualitative Analysis. Discourse analyses were conducted on the metaphoric
expressions, including the samples eliminated from the frequency analysis, em-
ploying the full context in which each sample occurred in the corpus. The purpose
was to reconstruct the conceptualization of anger for each 50-year period. Spe-
cific aspects of the conceptualization analyzed included the experiential scene that
is the basis of the concept, as well as the domains and mappings that instantiated a
metaphorical expression. Additional samples of particular features, collected from
the University of Virginia for the 15001849 period and from Cornell University
for 19th century British and American popular magazines, were used to aid in the
analysis of particular features of the metaphor samples collected from the Penn-
Helsinki and ARCHER corpora. These samples are presented and discussed when
appropriate to the discussion of the study results. The results of the frequency of
use and discourse analyses of the historical metaphoric expressions are presented
in Chapter 6.

An implication of the method: The value of mixed research designs

One peer reviewer commented that the study as described above is largely qualita-
tive with scant quantitative analysis. This is an important issue and needs to be
addressed. The CADS analysis method discussed in Chapter 1 is the methodolog-
ical basis for the current study, and this method specifically incorporates both
qualitative discourse analysis and quantitative measures. From a CADS point of
view, quantitative generally means the use of one or more statistical measure-
ments, either descriptive or inferential (or both). For the current study, the quan-
titative measure selected was the Normalized Frequency Rate (NFR). As discussed
previously, this measure was taken from the work of Douglas Biber and colleagues
in their corpus studies of lexical semantics. The NFR was specifically chosen for
the current study to calculate a standardized rate of lexical use, in order to com-
pare the rate of use found across the 490-year historical period under study and
reveal changes in that rate over time. The NFR has been in use for over two de-
cades by a variety of researchers in linguistics. It is a descriptive statistical measure
rather than inferential, but one that can be correlated with other data to reveal
Chapter 4. Research questions and methodology

important relationships between the factors under study. Most importantly, the
NFR fits the research questions of the current main study of historical metaphor.
As noted in Chapter 1, the challenges faced by researchers who choose to em-
ploy mixed study designs incorporating both qualitative and quantitative data
(such as CADS) are significant, but there is also potential for delineating impor-
tant phenomena that quantitative or qualitative studies alone may not discover.
We argue that mixed designs increase the power of the research study to reveal
such phenomena. The question concerning whether the design of the current
study is quantitative enough is left to the reader to decide; however, in our view,
the incorporation of quantitative and qualitative measures of any type in one study
is a distinct advantage of mixed research designs.

Chapter summary

The chapter presented an overview of the major issues that impinged on the design
of the current study, including the research questions and the study design for
both the ancillary study of non-linguistic data and the main study of historical
linguistic expressions of anger. The design issues included materials, data collec-
tion, and the data analysis procedures. The results of the ancillary study are pre-
sented in Chapter 5, and the results of the main study are discussed in Chapter 6.
Metaphor across Time and Conceptual Space

Table 1. The Historical Four Humors Texts: A Comparison of Basic Principles

Author Qualities Humors Organs Temperaments

Barrough, 1590 hot/cold blood liver melancholike

moist/dry phlegm spleen choleric
choler stomach others not
black bile mentioned.
Boorde, 1542 hot/cold blood Not mentioned sanguine
moist/dry phlegm specifically. phlegmatic
choler melancholike
black bile choleric
Bright, 1613 hot/cold blood liver sanguine
moist/dry phlegm Spleen phlegmatic
choler others not melancholike
black bile mentioned. choleric
Burton, 1621 hot/cold blood liver sanguine
moist/dry phlegm heart phlegmatic
choler gall melancholike
black bile spleen choleric
Charron, 1630 hot/cold blood liver sanguine
moist/dry phlegm heart phlegmatic
choler gall melancholike
black bile spleen choleric
Coffeteau, 1621 hot/cold blood liver melancholike
moist/dry phlegm heart choleric
choler gall others not
black bile spleen mentioned.
Cogan, 1605 hot/cold blood liver sanguine
moist/dry phlegm heart phlegmatic
choler lungs melancholike
black bile choleric
Cuff, 1640 hot/cold blood heart sanguine
moist/dry phlegm others not phlegmatic
choler mentioned. melancholike
black bile choleric
Dariot, 1598 hot/cold blood liver sanguine
moist/dry phlegm heart phlegmatic
choler gall melancholike
black bile spleen choleric
de Glanville, 1582 hot/cold blood liver sanguine
(translation of moist/dry phlegm heart phlegmatic
1360 ed.) choler gall melancholike
black bile spleen choleric
Chapter 4. Research questions and methodology

Author Qualities Humors Organs Temperaments

de Mediolano, 1609 hot/cold blood heart sanguine

moist/dry phlegm stomach phlegmatic
choler others not melancholike
black bile mentioned. choleric
Elyot, 1610 hot/cold blood brain sanguine
moist/dry phlegm heart phlegmatic
red choler liver melancholike
yellow choler stomache choleric
Huarte, 1698 hot/cold blood brain sanguine
moist/dry phlegm heart phlegmatic
choler others not melancholike
black bile mentioned. choleric
Lemnius, 1581 hot/cold blood heart sanguine
moist/dry phlegm brain phlegmatic
choler liver melancholike
black bile stomach choleric
Moulton, 1546 hot/cold blood heart sanguine
moist/dry phlegm liver phlegmatic
choler stomach melancholike
black bile choleric
Rogers, 1580 hot/cold blood heart sanguine
moist/dry phlegm liver phlegmatic
choler spleen melancholike
black bile gall choleric
Walkington, 1607 hot/cold blood heart sanguine
moist/dry water brain phlegmatic
choler others not melancholike
earth mentioned. choleric
Wright, 1601 hot/cold blood heart sanguine
moist/dry phlegm liver phlegmatic
choler brain melancholike
black bile choleric
chapter 5

Results of the ancillary study

of non-linguistic data


This chapter contains the results of the ancillary study of non-linguistic data de-
scribed in the previous chapter. The main analysis investigates the effects of the
five scientific advances on the cultural knowledge and practices of lay people dur-
ing the five-century historical period under study. In addition, the composite model
of the four humors model, developed from the 18 historical sources selected for
the ancillary study, is described in detail. The purposes of these analyses include
(1) to delineate the changes in cultural knowledge over time that may have affected
the conceptualization of anger and in turn, the meaning and use of linguistic met-
aphors of emotion; and, (2) to employ the composite model to aid in the analysis
of the samples collected from the text corpora for the main study of historical lin-
guistic metaphors of emotion.
Before discussing the results of the ancillary study, an important issue con-
cerning the necessity of this study must be addressed. A peer reviewer commented
that devoting an entire chapter to a secondary analysis seems unnecessary. The
methodological approach to research employed here is somewhat unconventional;
however, it is not unprecedented. As described in Chapter 1, the collection and
analysis of non-linguistic data has been available to linguistic researchers for many
years, and there are previous linguistic studies that have employed it, particularly
when investigating the effects of culture on language. The arguments of Lucy
(1996) and Enfield (2002) were discussed at length in Chapter 1 concerning the
methodological issues involving non-linguistic data and its effects on the analysis
of language samples, and ultimately, research results. These views were developed
in response to previous work in linguistics that had not properly taken into ac-
count the effects of culture on linguistic structure and meaning. Because the cur-
rent study is focused on the interaction between conceptualization and culture
and the effects of that interaction on language, the study is designed to take the
factor of culture into account via the use of non-linguistic data that provides infor-
mation on the cultural knowledge and practices of the historical period. For this
reason alone, the collection and analysis of non-linguistic data is necessary for the
current study design.
Metaphor across Time and Conceptual Space

However, the above explanation is not sufficient to justify the inclusion of this
chapter in the volume. Since the ancillary study is a secondary source of data that
is separate from the main study of historical linguistic metaphor of emotion, the
data collected and analyzed for the ancillary study could be inserted as needed
when discussing the analysis of the linguistic samples collected for the main study.
There are numerous points in Chapter 6 (on the results of the main study) where
ancillary data is included to support the analysis of the main study data. However,
including enough of the ancillary data to provide a clear and logical analysis for the
main study was seen as problematic: Chapter 6 would have been very long, and
due to the depth and breadth of the ancillary study data, some important points in
the analysis still would not have been presented clearly or in enough detail to meet
the goals of transparency and analytical rigor in scientific research.
Equally important, one of the main goals of the study the comparison over
time between the diachronic linguistic metaphor samples and changes in the Four
Humors model would have been much more difficult to present. This goal con-
nects directly to the research question concerning the relationship between
conceptualization and cultural models, so the comparison must present that rela-
tionship as clearly as possible. The comparison is best presented by providing an
analysis in the current chapter on the changes in the scientific and cultural model
of human health over the five-century period, followed in the next chapter by the
linguistic analysis of the changes in the meaning and use of the four keywords in
Chapter 6. By this procedure, the relationship between conceptualization and cul-
tural models can be viewed in much more detail; the reader is encouraged to keep
this comparison in mind when reading Chapters 5 and 6. For the three reasons
discussed in this section, the ancillary data is provided this chapter.

The ancillary study of the four humors cultural model

General principles of the four humors

Before discussing the ancillary study data across the five-century historical period,
two general features of the 18 historical Four Humors source texts need to be dis-
cussed. These features are the self-care focus of the source texts and the Macro-
cosm/Microcosm cultural model, termed in this volume the unified model.

The self-care focus

One interesting characteristic of all of the 18 Four Humors model texts (see
Chapter 4) is their common purpose as self-care books. Unlike professional works
Chapter 5. Results of the ancillary study of non-linguistic data

for scholars, the books were written for lay people to use in diagnosing and treat-
ing their own health needs and dealing with other important life issues. The med-
ical books in the group all have a self-treatment focus. Of the non-medical books,
several focus on a specific topic of human life that employed the Four Humors as
a method for providing insight on the topic. For example, Huartes (1698) Examen
de ingenios (Examination of Wits), argues that the Four Humors model can be
used to determine the best career path for an individual, based on an examination
of the persons temperament, as defined by the model. Huarte later gives specific
advice to women on what kind of man to marry in order to have intelligent chil-
dren, and to parents on foods that affect the intelligence and memory of a child.1
In another book (Dariot, 1598) on astrology, the Four Humors is employed to
show how to interpret the meaning of the position and movements of the stars and
planets to predict an individuals future prospects in work and marriage. The work
also takes the Four Humors one step further: in an appended article called Math-
ematicall Phisicke, the reader is shown how to diagnose and treat disease by using
astrological charts of the stars and planets in combination with the Four Humors
model. The charts were used to determine which times of a specific day and month
were best for specific humoral medical treatments. These works indicate that, dur-
ing the Renaissance, scholars applied the Four Humors model to areas of knowl-
edge outside of medicine in order to advise lay people on important life issues,
including career planning, raising healthy children, determining ones future po-
tential for success, and timing medical treatment for best effect. The principles of
the model were accepted by society as reflecting the realities of daily life as they
knew it, and the system was viewed as an insightful and practical body of knowl-
edge. The model was applied in sophisticated ways to current issues during the
historical period. Overall, the 18 sources cast the Four Humors as an integral part
of Renaissance popular culture.

The macrocosm/microcosm cultural model

Another common feature of the historical texts of the time deserves special men-
tion for its close relationship to the Four Humors. A cultural model of great influ-
ence in the Renaissance was the macrocosmos/microcosmos theory; Cuff (1640)
uses the terms the great world and the lesser world frames. The great world is the
universe, including the Earth, and the lesser world is humankind, including the
physical body. The model attempted to show that a human being is a microcosm,
or an imitative reflection, of the macrocosmic universe. To illustrate the relationship

1. Interestingly, Huarte recommended fish for this purpose, as many medical authorities do
today, yet for a different reason which fits the humoral theory.
Metaphor across Time and Conceptual Space

between the two worlds, Cuff created an analogy between the Sun and the human
And as in the midst of heaven there is seituated the Sunne, that enlighteneth all
things with his raies, & cherisheth the world, and the things therein contayned
with his life-keeping heat: for the heart of man, the fountayne of life and heat,
hath assigned to it by Nature, the middle part of our body for his habitation,
from whence proceedeth life and heat, unto all parts of the body, (as it were unto
Rivers) whereby they be preserved and enabled to perform their naturall and
proper functions (Cuff, 1640, p. 3).2

This passage describes in detail the concept of natural heat (e.g., life-keeping heat),
a feature of the Four Humors model of human health that was extended to natural
phenomena in the known universe, or the great world frame. The macrocosm/mi-
crocosm model is an extended analogy that established relational correspondenc-
es between the Universe and man. In Christian theology, both worlds were created
by God; therefore, in the Renaissance mind, it was logical to believe that, as the
work of one creator, both worlds must have been imbued with similar characteris-
tics. The correspondences between the macrocosmos and the microcosmos ex-
plained mans role in the universe and the natural processes and consequences of
life brought about by the influence of the greater world.
The macrocosm/microcosm model was also extended to the Four Humors
model. Cuff, a fellow of Merton College in Oxford whose book concerns the Four
Humors as it applied to different ages of human life (e.g., infancy, adolescence,
middle age, and old age), used the macrocosm/microcosm model to show other
similarities between the universe and humans that are related to the model. For
example, mortality was defined as the continual loss of heat and energy, both in
the natural world and in man, as a result of the aging process. The Renaissance
view was that man embodied known characteristics of the Universe, and therefore,
principles found in the natural world should also apply to human life and behav-
ior. The implications of these facts are discussed below.

Implications for the four humors model

A Unified View of the World. The intimate relationship between the macrocosm
and the microcosm led to some important consequences for the Four Humors
model. For one, the microcosm is significantly affected by events in the macro-
cosm; that is, the universe and earth control and dominate human life. This view
led naturally to studying the movements of the planets and stars, weather and cli-
mate, and geographical location to explain health, disease, intelligence, career

2. Spelling, punctuation, and italics are preserved from the original source text.
Chapter 5. Results of the ancillary study of non-linguistic data

prospects, and good or ill fortune in the life of an individual. For example, Burton
(1932/1621) lists three major types of causes for melancholy, a disease in the Four
Humors model which led to sadness, depression, and madness: the supernatural
(e.g., God and witches), the natural (e.g., astronomical events and negative life
events), and the body and its care (e.g., disease, poor diet, lack of sleep). These
categories theoretically include an infinite variety of specific causes of health, ill-
ness, and psychological conditions. The human body was in essence in intimate
union with the larger universe that is, the body was a moment-by-moment, one-
way reflection of the inherent characteristics, energy, and motions of the universe.
If an event in the greater world caused some sort of imbalance in that world, the
lesser world would show the effects of the imbalance.3
Thus, human life (the lesser world) in all aspects was a direct result of the pro-
cesses and events in the greater world, and the macrocosm/microcosm view was
combined with the Four Humors model to create a more powerful and insightful
model of human life. Combining the macrocosm/microcosm model with the Four
Humors created a unified theory of human development (hereafter, the unified
model) that explained how and why an individual had certain body characteristics,
personality traits, mental and emotional behaviors, skills and abilities, was fortu-
nate or unfortunate in life, and currently in good or poor health. The unified mod-
el was viewed as a powerful paradigm for explaining the impact of any life event
on an individual and also practical for making short- and long-term decisions.
An Open System. One other significant feature of the unified model was that it
was an open system; that is, the system could be entered at any level or point to
determine both the cause and result of any life event (Starobinsky, 1960, p. 9). For
example, from the point of view of the macrocosm, a recent change in seasons, a
weather pattern, or a star constellation could presage life events that could affect
health and fortune in the microcosm; actions of many sorts could be undertaken
to prevent the recent event from adversely affecting the person or improve its pos-
itive benefits. Conversely, a persons current medical condition, such as insomnia,
could be attributed to a wide variety of possible microcosm or macrocosm causes,
such as diet, another illness, personality type, the persons spiritual condition
(e.g., sin), the current season of the year, or a passing comet. Each of the possible
causes would be considered by a doctor or by the patient (in self-care), and each
would be eliminated in turn by employing the unified model until a cause was
identified, and then a treatment would be prescribed. If a single cause could not be
determined, two or more could be treated simultaneously. The unified model
could be entered at any level, from changes in the cosmos down to changes in

3. See later in this chapter, The Concept of balance in the Four Humors for more informa-
Metaphor across Time and Conceptual Space

bodily symptoms (and vice-versa) to aid doctors, clergy, and lay practitioners to
identify any cause of human happiness or suffering, and a course of therapeutic
action could be designed with a high degree of confidence within the model.


The self-care focus, the influence of the unified model, and the open nature of the
model have been discussed in detail for several reasons. First, the self-care purpose
of the historical texts demonstrates that the model was known and practiced by lay
medical consumers as well as trained doctors. This fact speaks to the wide dis-
semination of the Four Humors model at different levels of Western society, and it
implies the potential role of the model in influencing metaphoric language. Sec-
ond, the unified model had a direct effect on the Four Humors model; in the his-
torical texts, describing the latter model often invoked the former, either explicitly
or implicitly; therefore, the two models are inextricably intertwined and must be
studied as a single cultural model. Finally, the open nature of the unified model
created a system that was both easy to use by lay practitioners and complex enough
to be applied to any area of human life. These characteristics implicate the reasons
why the system was so popular during the Renaissance: the model was satisfyingly
insightful, straightforward in diagnosing health and mental issues, and applicable
to a wide variety of life concerns.
Based on the above considerations, the unified model was positioned at the
historical intersection of the Four Humors, fundamental scientific advances in
fields such as astronomy and biology, increasing lay interest in controlling per-
sonal health, success, and happiness, and the significant increase in knowledge
dissemination brought about by the invention of the printing press. The unified
model was a widely-shared cultural model of shared values and practices in early
Modern English popular culture. The ancillary study data was gathered on the
content and uses of the unified model during the historical period to determine
the basic principles and practices of the model. The data was in turn employed to
inform the analysis of the linguistic metaphor samples collected in the main study
of historical metaphors of anger.

The ancillary study of non-linguistic data: Results

This section combines three of the types of historical sources to describe the
unified model as it was constituted across the five centuries under study in the
main CM of anger study. There are two sections; first, the composite model is
Chapter 5. Results of the ancillary study of non-linguistic data

presented with a general overview of the unified model as described in the 18 his-
torical source texts. Second, information for the five scientific advances and the
changes to the model that the advances motivated are described.
In the scientific advances section, the collected information is presented for
each century in the following order: (1) details on the scientific advance; (2) details
on cultural practices related to the unified model of the time period; and (3) non-
metaphorical corpus data samples that explicitly reference the unified model. As
stated earlier in the chapter, the purpose is to develop a view of the model in each
time period and to delineate changes in the Four Humors model over time that
may have influenced metaphoric expressions and the conceptualization of anger.
Each century, with accompanying data from the ancillary historical study, is dis-
cussed in turn below.

The unified model: A historical composite view

The following is an overview (the composite model) of the major principles of the
unified, macrocosm/microcosm cultural model, as it was constituted and prac-
ticed in the 16th and 17th centuries. Included in the composite model are descrip-
tions of the five major principles discussed in the historical source texts: the four
qualities, natural heat, the four humors, the four organs, and the four tempera-
ments. The section begins with the basic principles of the model, followed by
principles specific to the macrocosm and the microcosm. Finally, based on the
findings, the implications for the current study are discussed.

Basic principles of the unified model

The model that will be presented here is a composite view compiled from the 18
historical Four Humors medical works.4 By this procedure, the tenets of the mod-
el that were widely accepted and disseminated via written texts were brought into
clearer focus. Five basic principles from the source texts will be presented in this
section. The first one, the four qualities, is related to the macrocosm, or the great
world of the universe and earth. The other four principles are within the micro-
cosm, or the lesser world of the human body: natural heat, the four humors, the
organs, and the temperaments (i.e., personality profiles). Each principle will be
presented in turn below.

4. See Table 1 at the end of Chapter 4 for an overview of the basic principles employed by the
18 historical source texts to describe the Four Humors model.
Metaphor across Time and Conceptual Space

The macrocosm principle

The Four Qualities. Each of the four qualities hot, cold, moist, and dry is always
paired with its opposite; the pairs are hot/cold and moist/dry. These were probably
proposed by Empedocles, a Greek philosopher of the 5th century B.C., according
to Ackerknecht (1982, p. 52). Empedocles envisioned that all matter came into be-
ing through the qualities; the four basic elements in the Greek system air, fire,
water, and earth were unique combinations of the four qualities. In turn, all mat-
ter (including the four humors; see next section) could be classified according to
combinations of the four qualities. The qualities are the basis, historically and the-
oretically, for the unified model.
An important aspect of the qualities which was discussed by many of the his-
torical Four Humors writers is the relative positive value of each pair to human
life. The qualities of heat and moisture were viewed as the most valuable of all be-
cause the Four Humors model held that these two were required for life. Without
them, living beings would die. Not surprisingly, cold and dry were viewed as less
beneficial, and at times, dangerous. The cold and dry qualities decreased heat and
moisture, and so in extreme cases led to sickness, detrimental changes in personal-
ity, and death. These two qualities were seen as useful in certain situations they
could be exploited by doctors to counteract excessive heat and moisture;5 how-
ever, in the end, cold and dryness had negative associations in the historical litera-
ture which were not generally ascribed to heat and moisture. Summarizing the
writers, heat/moisture was the life-sustaining pair of qualities, and cold/dryness
was the death-inducing combination.

The microcosm principles

Natural Heat. The first microcosm principle was mentioned in the previous dis-
cussion of the quote from Cuff (1640), on the relationship between the great world
frame of the universe and the lesser world frame of the human body. The body was
known to be perpetually warm, and this natural heat was thought to originate in
the heart:
...for the heart of man, the fountayne of life and heat, hath assigned to it by Nature,
the middle part of our body for his habitation, from whence proceedeth life and
heat, unto all parts of the body, (as it were unto Rivers) whereby they be preserved
and enabled to perform their naturall and proper functions (Cuff, 1640, p. 3)

5. In large quantities, all four qualities had negative effects on the body; see the section below,
The concept of balance in the Four Humors, for details.
Chapter 5. Results of the ancillary study of non-linguistic data

Like the sun in the great world frame, the heart was the source of life and heat in
the lesser world frame. This analogy explained both the ultimate source of life in
the body and the relationship between the human body and the world at large;
both were tied to and depended on heat to initiate and maintain living beings.
Heat also provided the energy for physical activity: changes in the level of heat
available explained changes in a persons level of activity. One factor influencing
the level of heat was advancing age in humans. For example, in the unified model,
a young person had a greater amount of heat relative to an older person, explaining
high activity in the young and low activity in the aged. Decreasing heat with age
also explained physical characteristics, such as rosy cheeks (the young) and white
hair (the old), as well as the presence of physical and mental health or infirmity.
In addition, regardless of age, the normal level of heat could vary from person
to person, affecting ones general temperament. For example, a high level of heat
could lead to either an outgoing temperament and/or a tendency toward violent
behavior; a low level of heat could lead to a penchant for solitude and/or lethargic
behavior. Natural heat was the most important of the microcosm principles, used
to explain a wide variety of physical, mental, and personality attributes.
The Four Humors. The second microcosm concept is the four humors. The hu-
mors are bodily fluids that were viewed as the most important for health. The four
fluids included blood, choler (sometimes termed red choler by the Renaissance
writers), black bile (typically termed melancholy) and phlegm. As was stated earlier
in this chapter, the Greeks associated the humors with the four qualities; each hu-
mor had two qualities, which were the same ones used to describe the four ele-
ments. Blood was hot and moist; choler was hot and dry; black bile was cold and
dry; and phlegm was cold and moist. Each person was dominated by one of the
humors, and so also took on characteristics of the two qualities associated with
that dominant fluid. One person was hot and moist due to the influence of blood,
another cold and dry due to black bile.
To counteract excess fluid, the four humors had to be in balance, or in equal
amounts. Balancing the humors required that the four fluids had to be of equal
proportion in the body to maintain good health. If a person had an excessive
amount of a humor, then disease would ensue. For example, an excess of choler,
the hot and dry humor, led to overheating the body with attendant symptoms and
disease; de Mediolano (1609) lists some of these problems, including ringing ears,
interrupted sleep and nightmares, upset stomach, little appetite, and overheating
(p. 22). The basic technique of humoral medical treatment was to mediate the
symptoms caused by the excess humor by applying treatments that had the quali-
ties opposite of that humor. For choler, moist and wet treatments would be pre-
scribed to counteract the effects of the hot, dry humor.
Metaphor across Time and Conceptual Space

Physicians used other methods to counteract excess humor and to restore

fluid balance; these techniques included diet, exercise, activities to maintain men-
tal health (e.g., socializing, music) and in some cases, cupping, in which excess
fluid (usually blood) was removed from the body. In the Renaissance, the concept
of balance was important to making life decisions, such as determining whom to
marry; selecting the proper mate would help to balance the humors optimally in
the couples children. Huarte (1698), for example, recommended that a hot and
dry man marry a cold and moist woman, in order to maximize their childrens
humoral balance. Such balancing would improve the childs health, intelligence,
and memory ability, according to Huarte.
Overall, the Renaissance writers viewed the effects of the humors on health in
relative rather than absolute terms. Due to the open nature of the Four Humors
model, a particular physical symptom could be the result of any of a large number
of causes, and these causes often occurred in complex combinations, reflecting the
complexity of life itself. Thus, an excess humor could cause a particular symptom,
or a humor could mitigate the effects of a symptom, or the potential effects of the
excess humor could be counteracted by another factor (such as weather or diet).
Several different causes could also occur in combination. The humors and the
other basic parts of the Four Humors model were viewed as possible causes in a
relativistic system of health.
The Four Organs. The third principle involved the organs of the human body.
In the original Greek humoral theory, physical organs, such as the heart or stom-
ach, were not included in the system, though they were known collectively as the
splanchma and played a role in the linguistic expression of emotion (Padel, 1992,
pp. 1314). This may have been the result of the scant knowledge that the
Hippocratic school had about the internal composition of the human body. The
techniques for examining the body at the time were few: visual inspection, touch,
and smell (Ackerknecht, 1982, p. 60). The body was seen as sacred, and there was
a strong prohibition against cutting open a dead human body for any purpose,
including science. In Classical Greece, the only accepted means to discover the
workings of the human body was to dissect the bodies of other animals. Until the
late Middle Ages, both Christianity and Islam continued to prohibit the dissection
of the dead.
Despite the lack of knowledge of human anatomy, the Four Humors model
included a detailed theory of human organs, and their effects on health and dis-
ease. There were four organs most commonly discussed in the historical texts, and
these included the heart, liver, brain, and spleen. In the Renaissance system, each
humor was associated with a specific organ: blood and the heart, choler and the
liver (or the gall bladder), melancholy and the spleen, phlegm and the brain. Due
to these associations between fluids and organs, the organs also took on the
Chapter 5. Results of the ancillary study of non-linguistic data

accompanying qualities. As a result, the heart was hot and moist (also incorporat-
ing the principle, mentioned previously, that the heart was the seat of natural heat),
the liver was hot and dry, the spleen was cold and dry, and the brain was cold and
wet. By the time of the Renaissance writers, the system had become much more
complex a physiological process of fluids moving between organs and through-
out the body. The process was described as follows.
When a person eats, food enters the stomach. The stomach breaks down the
food into basic nutrients, and the nutrient fluid (called chyle) was sent to the liver.
The purpose of the liver, according to the Renaissance writers, was to form the four
humors. Vicary (1577) explains the process of humor production.
Chyle which commeth from the stomacke to the lyver, should be turned into the
colour of blood...The naturals is sent with the blood to all parts of the body to be
ingendred and nourished. And the nutrimentals be sequestrate and sent to places
ordayned for some helpings. These are the places of the humours, the blood in the
Lyver, Choler in the chest of gal, Melancholie to the Spleen, Flegme to the Lungs
and the Junctures... (Vicary, 1577, p. 49).

Blood was formed first because it is the most important fluid for life, and it com-
prised the largest portion of the bodily fluid. Phlegm was next, followed by choler
and lastly black bile, in order of quantity and usefulness to the body. Blood was
most useful for life, and black bile least useful. In fact, the texts repeatedly pointed
out the dangerous nature of the melancholic fluid, due to its qualities of cold and
dryness. The writers agreed that black bile is only useful for helping digestion
(possibly due to the spleens proximity to the stomach). The heart then pumped the
naturals (or spirits, see next section) to the rest of the body via blood, and the
other three humors (the nutrimentals) were stored in their associated organs for
later use in helping digestion or restoring humoral balance. This process was re-
peated every time food was taken in.
Note that, in Vicarys account, the place of phlegm has changed from the brain
(in Greek thinking) to the lungs. To be more accurate, the Renaissance writers
were in some disagreement on the location of phlegm in the human body; of the
18 historical texts consulted, six linked the fluid to the gall bladder, five to the
brain, two to the stomach, and five others do not identify an organ or do not men-
tion phlegm. In any case, the writers agreed on the basic processes of nutrient
dissemination and humor production outlined above.
One other aspect of the organs to note was their ability, through the passing of
humors to the blood, and then to other parts of the body, to alter the natural heat
of the body. Temperature is a key concept underlying the health of the body in the
Four Humors system; the qualities of heat and cold were fundamentally important
to health, illness, personality, and length of life. In fact, several of the historical
Metaphor across Time and Conceptual Space

texts, such as Huarte (1698), argued that temperature was the key factor in areas
such as intelligence and career success. Again, like other aspects of the model,
temperature had a relative effect on health, but its role was well-defined in the
The Four Temperaments. Finally, we come to the concept that may have been
the most well-known and powerful feature of the Four Humors model the tem-
peraments. There were four types, including sanguine, choleric, melancholic, and
phlegmatic. These types, which were composite personality profiles, were based on
the concepts previously discussed, especially the four qualities and the four hu-
mors. The qualities determined basic categories which affected both the body and
personality. In addition, as mentioned previously, natural heat (body temperature)
was a key concept in health, and it was also important to individual personality: a
hot person often had red hair and an angry personality, and a cold person often
had white hair and exhibited behaviors similar to depression. The profiles were
logical extensions of the qualities that applied both to the physical body and to
Each temperament was seen as distinct from the others by the Renaissance
writers; however, two or more could combine in some individuals to cause physi-
cal and behavioral changes. Just as the four humors were viewed as having a rela-
tive effect on the body (rather than absolute effect), so the temperaments were seen
in terms of their relative influence on personality. Most people had one dominant
temperament which largely determined their stable personality traits, yet one or
more of the other types could affect behavior and emotions temporarily. These
changes in temperaments were caused by macrocosmic and microcosmic factors,
such as the stars and planets, the seasons, weather, geographic location, gender,
age, diet, and significant life events (e.g., a happy marriage or the death of a child).
In the unified view, virtually anything could affect personality over the short term,
and long-term change in behavior was possible due to permanent life changes,
such as advancing age. In short, the temperaments were seen as stable and distinct
in relative rather than absolute terms. The following are brief descriptions of each
temperament sanguine, choleric, melancholic, and phlegmatic including body
characteristics and behavioral traits.
1. The Sanguine Type. This type was viewed as the most prized of the four tem-
peraments. The sanguine type included the qualities of heat and moisture. In
terms of body characteristics, a sanguine person was characterized as having
red hair and skin, large veins, a good pulse, good digestion, and not prone to
disease; a few writers also included tall stature and/or large body type (though
not necessarily fat). In terms of personality, the sanguine type was cheerful,
kind, not susceptible to anger, social and outgoing, and loved entertainment,
Chapter 5. Results of the ancillary study of non-linguistic data

especially games and music. The sanguine type incorporated the highly posi-
tive, life-giving qualities of heat and moisture, and was often associated with
youth and the season of spring.
2. The Choleric Type. The choleric person was also hot, but dry rather than moist.
Bodily signatures included black hair and red or yellow skin, a very strong
pulse, a lean body type, unstable digestion (at least for hot foods, which
would increase the cholerics already high body heat) and difficulty in sleep-
ing. For personality, the major trait was a quick temper (all violent, fierce, and
full of fire; de Mediolano, 1609, p. 19), witty and bold in speech, proud, and
prone to fighting. These traits were commonly associated with men and the
summer season.
3. The Melancholic Type. Melancholic people were cold and dry, had dusky or
medium-dark hair and skin, a slow pulse, a very thin stature, poor digestion,
and typically were insomniacs. They were sad and depressed, had an anti-so-
cial tendency, were fearful and suspicious of others, loved solitude and quiet,
and enjoyed reading and quiet contemplation. Burton (1932/1621) stated that
students, professors, and clergy were prone to be melancholic, since their work
and interests are solitary by nature. The implication of greater intelligence re-
quired for these activities imbued the melancholic type with imagination and
a wry or sarcastic wit. Finally, middle age and autumn were often associated
with the personality type. It is also interesting to compare the melancholy per-
son to the sanguine type; they are by definition exact opposites, beginning
with their contrasting qualities (hot/moist vs. cold/dry). For this reason alone
the melancholic type was not favored in Renaissance society, due to its life-
sapping nature compared to the life-giving sanguine type.
4. The Phlegmatic Type. The phlegmatic person was cold and moist, had light hair
and skin, a fat, soft body (as opposed to muscular), narrow veins, weak pulse,
weak digestion, and slept heavily. The personality type may have been the least
appealing of all: dull in thought and speech (implying a lack of intelligence),
slow to respond and act, lazy, and showing little emotion of any kind. Women
were often thought to be phlegmatic, and winter was the common season as-
sociated with the type. The phlegmatic type was generally described in vague
terms in the historical texts, possibly due to its small number of distinguishing
features. Phlegmatic people were the anti-type, defined by the absence of vis-
ible signatures, rather than their presence.

The concept of balance in the four humors

The result of the Greeks original connection of the four qualities with the four
humors led to significant expansion of the Four Humors during the Renaissance.
Metaphor across Time and Conceptual Space

As one example, the concept of balance in humoral medicine, developed by the

Greeks, took on even more importance. Ackerknecht (1982) identifies three im-
portant ideas that contributed to the Greek idea of balance. These ideas include
(1) the fundamental tendency of nature (physis) to heal (implying that physicians
do not need to direct a cure but serve as an assistant during the healing process;
(2) eucrasia, the physical state in which the humors are in balance; and, (3) dyscrasia,6
an imbalance which leads ultimately to disease. These concepts reflected the Greek
goal of treating the entire human body holistically, rather than simply curing one
part (Ackerknecht, 1982, pp. 6162).
The expanded role for balance in the Renaissance model is likely an outgrowth
of the Greek-inspired symmetry between the humors with the qualities. The later
consequences included the development of specific advice concerning marriage
partners and the well-being of children, which extend the original goal of health
maintenance. Balance was a useful concept in the Four Humors system because it
could be manipulated by various means to improve health, natural abilities, and
other important aspects of life in the Renaissance. Many of the historical texts
specifically discussed balance via the need for moderation in all aspects of living,
neither doing too much or too little of any activity that would affect physical, men-
tal, or spiritual health. Even laymen understood the need for moderation. Matthew
Greens poem, The Spleen (Green, 1936/1737), is an example of the importance of
balance and the practice of moderation in lay medical practice that enacted the
principle, in the unified model.7

Five scientific advances in human physiology, A.D. 15001990

In Chapter 3, the study method included the identification of five major scientific
advances in human physiology to track the effect of scientific knowledge on meta-
phoric expressions of anger. The five advances were identified by consulting pres-
ent-day medical historians on the history of Western medicine. The advances
selected for the current study are shown below in Figure 10.
The rest of the chapter is organized by century in chronological order; each
historical period includes a detailed description of the major scientific advance in
human physiology, data from the study of the unified model cultural practices,

6. The concept of dyscrasia will have an important role to play in the results of the ancillary
study of historical metaphor; see later in this chapter on Rudolph Virchows discovery of cell
pathology in the 1800s.
7. See the section later in this chapter, The Eighteenth Century, for more details on the poem
and lay knowledge of the unified model in the historical period.
Chapter 5. Results of the ancillary study of non-linguistic data

15001599: Scientific anatomy (Andreas Vesalius, 1543)

16001699: Blood circulation (William Harvey, 1628)
17001799: Symptom localization (Giambattista Morgagni, 1761)
18001899: Tissue cell pathology (Rudolph Virchow, 1858)
19001990: Medical school standards (Abraham Flexner, 1910)

Figure 10. Five major scientific advances in human physiology, A.D. 15001990

and non-metaphorical corpus samples, followed by a discussion the effects of sci-

ence and cultural knowledge on linguistic expression of the period in question.
The summary section at the end of the chapter discusses several conclusions from
the analysis of the historical data.

15001599: Scientific anatomy (Andreas Vesalius, 1543)

The human body was traditionally seen as sacred; in many cultures, including the
Greek originators of the original Four Humors model, doctors and scientists could
not study the body in detail due to the ancient cultural prohibition against dissec-
tion. Due to this restriction, anatomical studies were conducted on other animals,
such as dogs, in order to understand physiology in general. The knowledge gained
from animal studies were then applied, by analogy, to human physiology.

Prior research

The most well-known of the early Western anatomists was Galen, who lived in the
second century A.D. His numerous studies of dogs, apes, and other animals were
read by many doctors and other experts in his own time, and the knowledge was
handed down from generation to generation with little revision or investigation
due to the cultural practice of regarding learned authorities (such as Galen) as the
final word on an academic topic. Eventually, Galens work was largely forgotten
during the Dark Ages. Then, in the Renaissance, his works were rediscovered,
translated from the original Latin to European languages, and published and dis-
seminated widely. Wear (1995) reports that in the 16th century alone, separate
editions of Galens writings were printed some 590 times (p. 253). Galens views on
human physiological processes influenced medical theory, practice, and physician
education greatly during the late Middle Ages and the Renaissance. Thus, Galen
had a significant influence on the Renaissance unified model theory, more than
1,200 years after his death.
The rediscovery of Galens ideas coincided with another significant change in
medical research: the long-standing prohibition against human dissection started
to change. By 1482, when the Pope formally granted permission for dissections of
Metaphor across Time and Conceptual Space

executed criminals to advance science, the practice had already begun in several
smaller states and kingdoms (Porter, 2002). The study of human anatomy through
the direct dissection of human corpses had a major impact on medical theory;
many of Galens views on physiology, based as they were on studies of non-human
animals, were refuted or significantly revised.

Vesalius work and influence

The first and most significant of these studies was conducted by Andreas Vesalius.
He took full advantage of the recent acceptance of human dissection and con-
ducted systematic studies of human cadavers, in order to write the first true hu-
man anatomy textbook, De humani corporis fabrica libri septum [On the fabric of
the human body in seven books] (1998/1543). Wear (1995) states that the book
marks a turning point in the medical view of the structure of the body (p. 275).
One of the main goals of the work was to investigate, using direct scientific obser-
vation, Galens views of the human body and its physiological processes. In this
effort, the book was a great success, and the work had several other important ef-
fects on medicine and science in general.
The influence of the book on medicine as a science included the establishment
of anatomy as the foundation of medical knowledge, the correction of many er-
roneous ideas about the body, due to the long-term reliance on Galens 2nd-century
A.D. dissection work on animals, and changes in the methods of investigation
turning away from analogical reasoning and toward direct observation of physio-
logical phenomena. Ultimately, this last change led to a significant decrease in the
influence of second-hand, learned authorities (such as Galen), and increasing em-
phasis on first-hand, visual inspection of physiological specimens and empirical
methods in science.
Vesalius accepted some of Galens views of the humoral model, including the
production of blood from chyle in the liver; however, Vesalius rejected many other
Galenic ideas, including the existence of two systems of blood vessels, one begin-
ning in the liver and the other in the heart. The effect of this correction is not obvi-
ous, but it would help inform the later discovery of the circulation of the blood
(see below). Overall, the many corrections of Galen, which Vesalius demonstrated
in his work, marked the beginning of the scientific refutation of the Four Humors
model, a process which would take 300 years to complete.

Evidence for the unified model in 16th century lay practice

Concerning evidence for the active practice of the unified model in the culture of
the mid-16th century, very little in the way of written data is available. The printing
Chapter 5. Results of the ancillary study of non-linguistic data

press was still a new invention, and books written for laymen were few and very
expensive. Andrew Boordes book Dyetary of helth (1542) was written for non-
experts and self-care purposes, and showed that the model was used by lay medi-
cal consumers, yet the book was probably read only by the literate upper classes.
The issue here is whether the Four Humors was known by a sufficient number of
the lay population to influence common cultural practice and language use. The
cultural evidence for the active practice of the model (that was found during the
ancillary study) principally comes from artistic works of the period.
Klibansky, Panofsky, and Saxl (1964) displays an extensive collection of 15th
and 16th century artistic works which incorporate principles of the Four Humors to
demonstrate to non-experts the basic tenets of the model. There are over 100 im-
ages in the collection, from many different artists, countries, and languages in
Europe. The use of pictorial representations would facilitate the dissemination of the
model to those members of society who were not educated and so could not read.
As a specific example, Roob (2005) includes an image of the period by
Thurneysser, painted in 1574. The four fluids are each represented in one quadrant
of a rectangle; in each quadrant, one part of an image of a person is displayed. For
example, in one quadrant a young womans clothed leg is shown; in another, the
left side of an older mans chest and head are displayed. The four quadrants join to
form a single, composite human being which possesses the major characteristics
of the four fluids and of human life youth and age, male and female. The zodiac
birth signs are also placed systematically in each quadrant, to show the relation-
ship between the Four Humors and astrology. Images such as these served as
teaching tools for the Four Humors across a society, and showed the active dis-
course and practical use of the model during the time of Vesaliuss work in the
16th century.
Linguistic evidence for the active use of the Four Humors in the 16th century
includes the following sample from a medical text in the Penn-Helsinki corpus,
written by Thomas Vicary, a physician who wrote medical treatment texts.
These are the places of the humors: the blood in the Lyuer, Choler in the chest of
gal, Melancolie to the Splen, Flegme to the Lunges and the Iunctures, the watery
superfluities to the Reynes and the Vesike. (1548).

The descriptions of the four humors, blood, yellow bile (choler), black bile
(melancolie) and phlegm (Flegme) and their locations in the four major organs
liver (Lyuer), the gall bladder (gal), the spleen (Splen) and lungs (lunges) fit a
typical Four Humors view of the human body, though there were minor disagree-
ments among medical experts at the time concerning which human organs com-
prised the four organs in the model. As discussed previously, some believed that
Metaphor across Time and Conceptual Space

the brain or the stomach was the fourth organ, not the lungs; for those experts,
phlegm originated in the brain or the stomach.
Below is another example from the Penn-Helsinki corpus, written by William
Clowes, a surgeon:
First as I said, euacuation going before, to diminish the humors sore abounding, it
was therfore thought most meete to begin with bloodletting in the middle vain on
the left arme, & I did then take from ech of them vii. or viij. ounces of blood. The
next day following they were also well purged with this purgation, R. Diasenae
z. j. ss. Sirr. fumariae, z. j. Aquae scabiosae, z. iij. Misce. and herewith they were
purged. (1596)

The text displays several types of treatments popular in the unified model, includ-
ing bloodletting or cupping (the opening of a vein to remove excess bodily fluid,
usually blood), evacuation (i.e., purgation), and the use of several medicinal plants.
During the 16th century, Vesalius scientific work in human anatomy did not have
any discernible effect on the knowledge and practice of the unified model.

16001699: Blood circulation (William Harvey, 1628)

Harveys work and influence

The discovery of the circulation of the blood had a profound effect on scientific
understanding of the body. Ackerknecht (1982) calls Harveys (1958/1628) work
the greatest physiological advance of the seventeenth century, and perhaps all
times... (p. 113). The discovery was called circulation because of the circular mo-
tion of the blood, in a clockwise path around the body.
Important parts of the Renaissance Four Humors theory were refuted by this
scientific advance, including Galens view that bodily fluid flowed upward from the
intestines through the liver and heart to the brain. Harvey had a significant impact
on many aspects of human physiology, though the effect was not immediate.
Ackerknecht (1982) reports that opposition to Harveys circulation theory was
strong; Harvey himself published a series of papers answering the counterargu-
ments of his critics, including several written in response to John Riolan, an anato-
mist and pathologist. The full effect of Harveys work would not be realized for
many years.

Evidence for the unified model in 17th century lay practice

Evidence for cultural practices of the unified model in this period includes a grow-
ing number of scientific treatises written for laymen which analyzed a plant or
other natural substance according to its humoral qualities. One example is a
Chapter 5. Results of the ancillary study of non-linguistic data

pamphlet written by C.T. (1615) entitled, An advice how to plant tobacco in

England. The purpose of the work was to discuss in detail the virtues of the tobacco
plant and its medical uses; after these advantages were established, an extended
account of the method for planting and cultivating the plant was described.
The unified model was employed as an analytical tool for describing the ad-
vantages and uses of tobacco. C.T. provided several reasons for the popular use of
tobacco among the Indians in South American and also the Spanish, and some of
these reasons are directly related to the model. For one, the plant opens the body,
and lets out heat by the pores. This was important in the Four Humors because
closed pores led to overheating and the attendant medical problems, including ill-
nesses associated with the gall bladder and its hot and dry humor. Also, tobacco
was useful for drying excess moisture, a characteristic that was useful for unified
model medicine because an overabundance of moisture was seen as damaging to
health, just as excess heat was injurious. A third reason advanced by C.T. for using
the tobacco plant was that it cured dropsie (humoral fluid seeping out of the heart,
thought to be a cause of heart failure). These statements are also evidence that lay-
men bought and used the pamphlet for the self-treatment of illness within the
unified model. In addition, the pamphlet lends support to the active use of the
unified model during the historical period of Harveys study of circulation.
Further evidence of the use of the unified model in practical science includes
another treatise written by Henry Stubbe (1662) called The Indian nectar, or, a
discourse concerning chocolata. Cacao nuts had been recently imported from the
New World (Brazil), and Stubbe performed an analysis, using principles in the
unified model, concerning the uses of chocolate as a healthful drink. Many med-
ical terms are used, such as the proper dose to be taken, and referring to choco-
late as a compound. In addition, Stubbe makes use of unified model terminology
when describing the health effects of chocolate: yields good nourishment to the body, it helps to digest the ill humours, void-
ing the excrements by sweat, and urine: and I say, it is no where more neces-
sary then [sic] in the Indies, which are moist, and apt to create lassitudes, their
bodies there being, together with their Stomachs, full of Phlegm, and superflu-
ous moisture, which are concocted by the heat of Chocolata into good Blood...
(pp. 8586).

Scientific research of the 17th century actively employed the unified model to in-
vestigate practical issues in everyday life, including the health effects of new plants
like tobacco and cacao nut.
Finally, data samples from the Penn-Helsinki corpus also corroborate the ac-
tive use of the unified model. Letting blood is the common term for cupping, the
Metaphor across Time and Conceptual Space

unified model practice of removing excess humor as a treatment for various ill-
nesses. The sample shows the active use of the practice by laymen.
...and yesterday morning I sent fore a curgen at Bischops Castell, that let Mrs.
Wallcot blud, and he pricke my arme twis, but it would not blled; and I would not
try the third time. (1633)

Another example of bloodletting shows the relationship between the cultural

practice and specific illnesses treated via the unified model.
Now for that his Ulcers were many, and subiect to a hotte distemper, for that cause
hee might the better admitte bloud letting, being also a man of a growne age,
therefore I tooke the more quantity thereof. (1602)

In the sample, the author, a medical doctor, uses bloodletting to alleviate the symp-
toms of ulcers, which included a hotte distemper.
To summarize this section, Harveys discovery of blood circulation was a ma-
jor advance in physiology (and a major blow to the unified model), yet there was
no immediate effect on the practice of the unified model among doctors and lay
medical consumers.

17001799: Symptom localization (Giambattista Morgagni, 1761)

Morgagnis work and influence

A major scientific advance was the publication of De sebidus et causis morborum

(On the Sites and Causes of Disease) by Giambattista Morgagni (1960/1761), a five-
volume work which advanced the theory of local, clinical symptoms of disease,
rejecting the holistic, global view of bodily changes described in the unified mod-
el. The work summarized the results of 700 human autopsies, including analysis by
microscope, a recent invention; the number of autopsies allowed for generalizing
the results for the field of human physiology. Specifically, Morgagnis results ...
demonstrated that diseases are located in specific organs, that disease symptoms
tally with anatomical lesions, and that pathological organ changes are responsible
for most disease manifestations (Porter, 1995, p. 410).
The immediate effect was the corroboration of Vesalius earlier work in anato-
my with the addition of finding the sites of disease not in fluids but in bodily tis-
sues. The work also heralded the growing use of the microscope for physiological
study, ultimately leading to Rudolph Virchows seminal study on cell pathology
(see the next section). Morgagnis study showed that the unified model was near
the end of its run as a viable description of human body processes.
Chapter 5. Results of the ancillary study of non-linguistic data

Evidence for unified model in 18th century lay practice

Yet, the unified model was still practiced among doctors and lay medical practitio-
ners alike during the period. Cultural evidence for the continuing practice of the
unified model includes a field guide used by physicians when treating illness, a
book-length poem, popular drinking songs published anonymously in the 1770s,
and linguistic data from the ARCHER corpus. These are discussed in turn.
The field guide, titled A pocket dictionary of medicine, midwifery, and surgery
by Matthew Wilson (1787), gives brief descriptions of treatments for various ill-
nesses. For example, the treatment for rheumatism was given as follows.
Bleed, & give a mercuri vomit Sweat [with] Gum Guiai...if no inflammation, rub
in [or?] Flesh Brush Volatile or Saponaceous Linement (no page number).

The treatment includes several techniques developed in the unified model, includ-
ing cupping (bleeding) and vomiting (a purgation procedure to evacuate illness-
causing material from the body), as well as checking for inflammation, or heated
fluid, which resulted in fever. Wilson also recommends taking cold baths of salt
water, in order to reduce the incidence and severity of fever; a cold treatment to
counteract hot fever follows the unified model principle of counteracting the ill-
ness-causing effects of one of the four qualities with its opposite.8
The poem, called The Spleen by Matthew Green (Green, 1936/1737), a busi-
nessman, listed in verse many types of activities which affect the balancing of
black bile. Activities which Green recommended for promoting positive balance
included food without excessive seasoning, exercise, merriment, entertainment
(including plays and concerts), reading, social places like coffee-houses, social
events (including the company of women), and an outgoing personality. Con-
versely, Greens list of activities which cause imbalance included lawsuits, gam-
bling, passion (i.e., extreme emotion), party politics (including reforming
schemes), financial ventures, fanaticism of any kind, and superstition. These lists
show that health was viewed as a reflection of a persons total life and activity, in
agreement with the unified theory, and also that non-professional lay practitio-
ners were aware of the importance of balancing the humors for physical and men-
tal well-being.
The drinking songs, from a tavern songbook published by William Jackson
(1770), includes several which allude to principles of the unified model. For ex-
ample, one song refers to symptoms of melancholy.

8. See earlier in this chapter for related information: The Concept of balance in the Four
Metaphor across Time and Conceptual Space

Whilst from our eyes fair nymph

You guess the Secret passions of our mind
My heavy eyes you say confess
A heart to love and Grief inclind.
Heavy eyes refers to sadness, both symptoms of melancholy; the last line (Grief )
also refers to the condition.
Another song refers to the principle in the unified model to prescribe treat-
ments (including activities) which counteract the effects of melancholy.
Mirth and Humour do unite us
Joyful songs will merry make us
Melancholy will Forsake us.
Though to the contemporary mind making merry seems less a prescription than
everyday common sense to cope with sadness, historically such activity was a for-
mal medical principle. As was discussed in the previous section on the unified
model, the model had multiple levels, including the human body, the physical
world, the heavens (stars, planets, and celestial events like passing comets), and the
spiritual world. Part of the human body was a persons current temperament; if a
negative view of the world and life was experienced by the person, then sickness or
mental disease was the medical result. The sample above shows that melancholy
(i.e., excess black bile) was removed by laughter and having fun. Matthew Greens
poem, The Spleen, discussed previously, also stated that merriment restored the
humoral balance.
Finally, linguistic samples from the ARCHER corpus include the following
from a private diary, dated 1720.
Mar. 14. After some vellications & preludes the Gout seizd
upon my right foot in the bones of the Tarsus. I let blood &
found it very much inflamd, & laid a Caustic upon the part,
drinking much water & sugar & juice of lemon, fasting, & taking
aloes every day. I made a crucial incision & causd an issue
when the Caustic was laid.
In the unified model, bloodletting or cupping was used for many different types of
physical symptoms, since heated blood brought on illnesses of many kinds. The
drinking of water with sugar and lemon juice was meant to cool the blood and
symptoms of gout. Fasting was used to purge foods in the diet that may have caused
the inflammation.
Chapter 5. Results of the ancillary study of non-linguistic data


Based on the data described in this section, there is evidence for the cultural prac-
tice of the unified model during the 18th century period. The model was still in
active use among lay medical consumers.

18001899: Tissue cell pathology (Rudolph Virchow, 1858)

In the early 19th century, medicine was enjoying the fruits of the many scientific
advances of the 17th and 18th centuries, including the development of the micro-
scope, discussed in the previous section. However, the Four Humors was still ac-
tively used in medical practice and in scientific theories of human physiological
processes. One of the important humoral concepts practiced in the 19th century
was dyscrasia. Dyscrasia was one of the original principles developed in the
Hippocratic school, which stated that humoral imbalance was the cause of disease
(Ackerknecht, 1982, p. 6162). The concept was still in use among pathological
anatomists (investigating disease via dissection), including Carl Rokitansky of the
New Vienna School. The New Vienna School was known for its objective accounts
of disease, eschewing the influence by any particular theory; yet, in 1846, Rokitansky
published a book entitled the Handbook of General Pathological Anatomy, which
argued that dyscrasia in bodily fluids was the cause of disease. Humoralist princi-
ples still influenced the scientific study of anatomy and physiology of the time.

Virchows work and influence

Rudolph Virchow, a pathologist in Berlin, advanced a new view which refuted the
dyscrasia theory, and ultimately humoralism as a scientific model of human phys-
iology. There were two major forums through which Virchow proposed his theory.
First, he wrote a review of Rokitanskys book which completely demolished the
dyscrasia theory, and Rokitansky as a result deleted the theory from later editions
of his book (Ackerknecht, 1982, p. 166). Second, in 1858, Virchow published a
book of his research, Cellular Pathology, which again denied that bodily fluids
were the central cause of disease; his theory, for the first time, placed health and
disease at the level of cell tissue.
Virchow used many examples of cells from microscope investigations, as well
as discussions of known illnesses, to support his view. For example, the dyscrasia
view would predict that disease persists, as Virchow states, in the blood itself ;
that is, bodily fluid is a substance that can propagate disease independently with-
out the help of other parts of the body. One of Virchows arguments to refute this
prediction involved the effects of alcohol on the blood. It was accepted fact at the
Metaphor across Time and Conceptual Space

time that drunkenness was not a permanent condition, but required additional
and regular intake of alcohol to be maintained; Virchow argued against the hu-
moral view of permanent dyscrasias in the blood on this basis. He counterargued
that every dyscrasia is dependent upon a permanent supply of noxious ingredients
from certain sources (Virchow, 1940/1858, p. 131, italics mine). Virchow thus
argued that the local origins to account for a disease are found in the bodily tis-
sue and organs. With similar arguments, he convincingly refuted the dyscrasia
view, and with it, the last remnants of the unified model. Virchows research and
writing changed scientific research and medical practice significantly in the sec-
ond half of the 19th century.

Evidence for unified model lay practice in the 19th century

The evidence for the active practice of the unified model during the nineteenth
century includes a list of physicians fees and services, and also language samples.
First, the fee list, published by the Portsmouth, New Hampshire Medical Society in
1806 (Estes & Goodman, 1986, pp. 3033). The list shows treatments commonly
used in the Four Humors model, such as bleeding (the removal of excess blood);
making a seton (a drain for fluid below the skin); paracentesis (the removal of
fluid from the chest or abdomen); trepanning (opening the skull to relieve fluid
pressure on the brain); and the glyster syringe (for administering an enema, which
treated constipation but also Four Humors stomach ailments).
Second, language samples from a popular magazine of the nineteenth century,
Littles Living Age, from the Making of America corpus at Cornell University, also
show the practice of unified model was active at the time.
...who can venture to doubt that enlargements of the liver, affectations of the
spleen, hypochondria, jaundice, and gout, with sundry other maladies less
admissible into our pages, will be effectually softened down, washed away, and
expelled. Who can be surprised that during the ten years that these wonder-
working waters have been flowing, the city (!) of Homburg has greatly improved...
(1852, Volume 32, Issue 403, p. 257)

The passage makes reference to the effect of hot mineral springs on affectations of
the spleen. In the unified model, diseases were treated with substances which
were believed to have the opposite qualities of those that cause the disease. For
example, for symptoms which were the result of black bile (the cold and dry hu-
mor), hot and wet treatments were applied. Therefore, the hot springs of Homburg
were viewed as an effective treatment. The passage was written before Virchows
book was published, so his discovery has not begun to affect language use.
Chapter 5. Results of the ancillary study of non-linguistic data

As was seen in previous historical periods, new knowledge takes time to in-
corporate into medical and cultural practices. Another passage from Living Age
later in the century, almost 20 years after Virchows discovery, also alludes to the
continuing value of the unified theory in the nineteenth century.
And he who had been so exactly easy and affable to all men that his face and coun-
tenance was always present and vacant to his company, and held any cloudiness
and less pleasantness of the visage a kind of rudeness or incivility, became on a
sudden less communicable, and then very sad, pale, and exceedingly affected with
the spleen. (1877, Volume 131, Issue 1696, p. 196)

The sample implicates a medical relationship between the spleen and sadness. The
connection between the organ and physical/mental health is clear, and the signs of
illness described follow the tenets of the unified model: low social activity, depres-
sion, and a change in skin color.


The medical society fee schedule and the language samples show that 19th century
medical treatment and cultural practice actively applied unified theory methods,
up to two decades after Virchows recent discovery that disease originates in cell

19001990: Medical school standards (Abraham Flexner, 1910)

Medical training for doctors in the United States in the early 20th century was in a
state of flux. In the latter half of the 19th century, state medical boards were being
established to create and enforce high standards of medical practice and also ethi-
cal behavior for medical professionals. Apprenticeship was still the common
means for training students; however, the quality of the graduates was highly vari-
able. Careers in medicine were booming as the countrys population grew, creating
new opportunities for the profession. Non-profit medical schools, such as Johns
Hopkins and Harvard, had been established in the early part of the 19th century,
and provided a quality education. In the latter half of that century, some of the new
schools which opened to meet the rising demand were for-profit institutions, and
these usually had lower standards for both admissions and graduation and poorer
teaching facilities and equipment. The issue of training for doctors quickly became
an important topic for the medical community. Ludmerer (1985) states that
ter 1900 a broad consensus began to appear in the medical profession regarding
the desirability of improving medical education (p. 168). This consensus also
Metaphor across Time and Conceptual Space

coincided with the rise of scientific research in medicine in the 19th century, based
on first-hand observation rather than relying on second-hand models such as the
Four Humors. The American Medical Association, created 50 years earlier as the
national professional society, devised a plan for dealing with the increasing con-
cerns for medical education.

Flexners work and influence

In 1904, the AMA created an internal group, called the Council on Medical Educa-
tion, to study educational reform. The council in turn proposed to conduct a na-
tionwide survey of all 168 medical schools currently existing in the U.S. After a
preliminary survey in 1906 confirmed the need for a detailed investigation, the
Council requested that the Carnegie Foundation for the Advancement of Teaching
perform an independent effort. The Foundation hired Abraham Flexner, an edu-
cator who had developed a detailed philosophy of education which included a
concrete, experiential (i.e., learn by doing) component. In 19081909, Flexner
traveled to all 168 medical schools currently operating to gather information. The
final report he wrote to summarize the findings is called the Flexner Report
(Flexner, 1910).
The Report covers a wide array of issues in medical schools, above and beyond
the classroom pedagogy. Part I includes historical information on medical educa-
tion in the States, coursework, standards for school finance, the effect of medical
sects (such as homeopathy and osteopathy) on training, the role of state medical
boards and postgraduate schools, and the issues of training for women and
African-Americans. Part II is a listing of all 168 schools surveyed, organized by
state. The Appendix contains a table summarizing the collected information for
each school. Flexners conclusions concerning the state of medical education in-
clude the importance of teaching the principles of scientific inquiry, the advan-
tages of original research in the school, the necessity of hands-on learning, the
requirement of proper credentials for the teaching faculty and minimum educa-
tional standards for admitted students, and the problems posed by the medical
sects. The original report is 346 pages long and covers each topic in detail. The
Report was influential in the years immediately following its publication; standards
for medical school education were instituted and enforced in the States through
state licensing of the schools (Beck, 2004). A recent article which surveyed the
original 168 schools (Hiatt & Stockton, 2003) found that 12 had closed or merged
with other institutions within a decade of the Report, and 26 more closed or merged
within two decades (p. 37).
Chapter 5. Results of the ancillary study of non-linguistic data

Evidence for unified model lay practice in the 20th century

The issue of medical sects is of interest for the current study. Flexner discusses this
topic in Chapter 10 of the Report. The sects he refers to are the three main theo-
retical perspectives then currently in vogue among doctors. These include allopa-
thy (regular medicine), homeopathy, and osteopathy. However, these last two are
descendants of earlier sects which had existed in the 19th century. These older
sects were established due to two problems of medicine in the early half of the
1800s: the lack of clinical training for new medical students, and the public rebel-
lion against regular medicines use of humoral practices that came to be viewed as
extreme measures, such as bloodletting and purging. These two problems caused
the lay public in the 19th century to distrust doctors and to turn to new systems
which promised effective treatment without the use of the unpopular humoral
treatments (Rothstein, 1972).
Though homeopathy and osteopathy were the last remnants of the old sects,
some medical schools still taught some of the old humoral practices, including the
Thomsonian and eclectic sects. These two were botanical sects, in that they pro-
moted the use of herbs and other plants for maintaining health and treating illness.
The major principles of Thomsonian medicine rested on the preservation of heat
in the body and the elimination of coldness, similar to the unified models four
qualities. Eclectic medicine depended on medicines which were emetics used to
induce vomiting, and cathartics to empty the bowels; both of these were also simi-
lar to components of the older humoral system. Thus, some medical colleges of the
early 20th century included instruction in the medical sects which employed ele-
ments of the unified model. When the Flexner Report was enacted by the AMA to
reorganize medical education, these theoretical systems were no longer permissi-
ble in the teaching colleges. As a result, after 1910 the unified model was finally
eliminated from both professional medical practice and the training of physicians
in the United States.
In addition, a 1925 newspaper story from the New York Times provides evi-
dence for cultural knowledge of the unified model in the 20th century. The story
reported that a Cambridge University professor, Barcroft (no first name given), is
at present analyzing the spleen where people are popularly supposed to keep their
bad tempers (December 19, 1925, p. 10). The reference to the societal belief was
an indication of the value of the unified model in popular culture of the time; how-
ever, direct evidence for the practice of the model was not found in any source for
the historical period. In addition, the statement did not indicate any evidence of
knowledge of the blood and heart as sources of anger and desire; all negative emo-
tions are attributed to the spleen, which was a significant divergence from the view
of the historical unified model. Knowledge of the model had decreased since the
Metaphor across Time and Conceptual Space

1850s and what remained reflected broad themes concerning the relationship be-
tween the human body and emotions, rather than the detailed knowledge found in
previous centuries.

Implications for the current studies

There are several implications of the ancillary study of non-linguistic data for the
main study of diachronic metaphors of anger. First, the major identifying fea-
tures of the unified model include the systematic, all-inclusive categorization of
phenomena that are external to the body as well as internal; the close, causal con-
nections between these two realms of experience; and, the explanatory power of
the model to explain any event, whether supernatural, natural, physical, or psy-
chological across the human life span. The systematic and inclusive nature of the
model creates a unified theory of human development, tying together all known
realms of the macrocosm/microcosm as understood in the early Renaissance.
Second, the unified model was insightful and powerful as an explanatory tool.
Modern characterizations of the theory tend to summarize its major parts and to
gloss over the details. Though the review in this chapter does not claim to be ex-
haustive, in the historical texts there is a great depth of thought and explication of
the model. As a result, the system was able to explain and provide advice for all
important life issues, including healthy lifestyles, mental health, considerations of
career choices and marriage partners, child-rearing, and the effects of cataclysmic
events on the quality of life. The model was seen by the historical writers not only
as a scientific theory, but a practical reflection of reality, useful for making short-
and long-term personal life decisions.
To address these needs, some authors combined the unified model with other
theories, such as astrology and mathematics, to create specialized decision-mak-
ing systems that shared in the practical explanatory power of the unified model. In
practice, the theory was straightforward, adaptable, and useful in early Modern
English society. The large number of publications on the topic in the 16th and 17th
centuries testifies to the high cultural value ascribed to the Four Humors and im-
plies the wide distribution of information about the system in different levels of
society, further suggesting that the unified model influenced language and the
cognitive conceptualization of the human body.
Third, the unified model included concepts that parallel those found in Lakoff
and Kvecses (1987) analysis of the CM of anger. Heat, pressure, and visible
body symptoms (skin redness, agitation) were all found in the descriptions of the
major tenets of the Four Humors, especially in the four temperaments. The san-
guine and choleric types in particular showed important parallels with the folk
Chapter 5. Results of the ancillary study of non-linguistic data

theory. For example, sanguine people were described as passionate and show emo-
tion easily, though they also controlled the negative effects of anger well. Choler
was described as heating the blood, leading to sweating, skin redness, anger, and
violence, and the lack of control over emotion is apparent in the descriptions of the
choleric personality. The melancholy type was also described in ways that were
similar to Lakoff and Kvecses non-prototypical anger cases, particularly the con-
cepts of low heat, fluid pressure, and the ability to maintain emotional control
through the periodic release of black bile. These apparent similarities between the
historical unified model and Lakoff and Kvecses theory informed the research
questions and the design of the current study on anger metaphors in historical
culture. Therefore, the connection between these conclusions and the non-proto-
typical cases of anger discussed by Lakoff and Kvecses (including He vented his
anger and He vented his spleen) suggest the need for further investigation.
Finally, the data gathered for the ancillary study showed that the unified mod-
el was actively practiced by lay medical consumers during the historical period, at
least between the early seventeenth century and the late decades of the nineteenth
century. The scientific advances that refuted important parts of the model during
this time period did not seem to have discernible effects on lay medical practices;
the evidence suggests that the unified model was an important influence on medi-
cal self-treatment by non-experts throughout the historical period under study,
and its effects may have stretched beyond medicine into metaphoric language, and
eventually to cognitive conceptualization. This issue is the investigative goal of the
main study of diachronic metaphorical expressions of anger; the results of that
study are presented in the next chapter.
chapter 6

The main study of two diachronic

metaphors of anger


The metaphoric expressions collected from the Penn-Helsinki and ARCHER cor-
pora will be analyzed over the course of the five-century historical period. The
findings from the previous chapter, the ancillary study of non-linguistic data, will
be employed to aid the interpretation of the linguistic samples.
The chapter is organized in sub-sections. First, the keyword data collection
results are summarized. Second, the frequency tables of the metaphor samples are
presented and the results across the historical period, given in 50-year increments
(or cells) are discussed. Third, changes in the structure, meaning, and use of the
metaphoric expressions for each 50-year cell are analyzed. In addition, the fre-
quency over time is compared to the five central scientific advances (identified and
described in Chapter 5), in order to investigate the relationship between changing
knowledge of the human body and its physiological processes, (both expert and
lay knowledge), and metaphor. Finally, the results of the study are summarized at
the end of this chapter.

Data collection results

The keyword search totals are provided in Table 2.

A total of 62 metaphor samples were collected from the corpora. However,
several had to be eliminated in order to calculate the NFR (see Chapter 4). Recall
that the two corpora overlap between the years 1650 and 1720; using all of the data

Table 2. Corpus keyword data collection results

Corpus Total Total emotion Overlap Unclear Total cases

Keywords Metaphors cases cases for analysis

Penn-H 308 28 3 0 25
ARCHER 86 34 7 2 25
Totals 394 62 10 2 50
Metaphor across Time and Conceptual Space

collected in this period would artificially inflate the NFR calculations for the
16501699 and 17001749 periods. To eliminate the overlap, a total of 10 cases
were deleted from the two corpora: three cases from the Penn-Helsinki (hereafter,
P-H) dated between 1700 and 1720, and seven cases from ARCHER dated be-
tween 1650 and 1699; a total of 52 cases remained. After the elimination of the
overlap cases, the P-H covers the years 1500 to 1699, and ARCHER covers 1700 to
1990. Finally, two cases were eliminated because the structure and contextual
meaning did not clearly instantiate a conceptualization of anger or any other
emotion, leaving a total of 50 metaphoric expressions for analysis.

The frequency results

Table 3 summarizes the emotion metaphor frequencies for the keywords for each
50-year time period between 1500 and 1990 in the two corpora. The raw frequen-
cy (total instances per period) was also converted to the Normalized Frequency
Rate (hereafter, NFR) of one instance per 1.0 million words of running text in each
corpus. The NFR was calculated for each keyword to provide a means to compare
the use of each keyword across the entire five century time period. The last year
that a keyword appeared in a corpus is shown in parentheses.

Table 3. Metaphor frequency counts, total and by keyword, A.D. 15001990

Year Range vent- spleen blood boil- Raw NFR


15001549 0 0 0 0 0 0.0
15501599 0 0 3 0 3 1.7
16001649 1 3 4 1 9 5.0
16501699 2 2 5 1 10 5.5
17001749 1 4 2 0 7 4.1
17501799 4 0 3 0 7 4.1
18001849 1 0 4 0 5 2.9
18501899 2 0 0 3 5 2.9
19001949 0 0 0 1 1 0.5
19501990 0 0 0 3 3 1.7
TOTAL 11 9 21 9 50 XXX
Chapter 6. The main study of two diachronic metaphors of anger

General trends

Several general historical trends are delineated in Table 3. First, the raw frequen-
cy of the metaphoric expressions increases over time during the 15001699 pe-
riod, from zero cases in 15001549, three in 15501599, nine in 16001649, and
10 cases in 16501699. The high point of the raw frequency data thus occurs in
the late 17th century, which was also the high point in popularity of the unified
model, in terms of the number of books published on the topic (see Chapter 5).
During the 18th and 19th centuries, the raw frequency decreases over time, from
seven cases in the two 50-year cells in the 18th century to five cases in the two
cells in the 19th century. In the 20th century, the lowest frequency of one case
occurs in the 19001949 cell, and then there is an increase to three cases in the
1950 to 1990 cell; interestingly, the keyword boil- comprises all four cases in the
20th century.
The NFR begins at 0.0 between 1500 and 1549, increases to 1.7 between 1550
and 1599, then increases to 5.0 in 16001649 and then reaches a high of 5.5 in the
16001699 period. From that point, the NFR gradually decreases, to 4.1 in the
18th century, 2.9 in the 19th century, and finally to a low of 0.5 from 1900 to
1949. In the last 50-year increment (1950 to 1990), the NFR increased to 1.7.
Since the selected corpora are representative of English usage during the histori-
cal period under study, the trends indicate the general patterns of use of the
keywords in metaphoric expressions by native speakers during the 490-year his-
torical period.
In addition, the use of each of the keywords varies in raw frequency over time;
however, each keyword exhibited a different diachronic pattern. The noun blood
occurred the most often, 20 times between 1550 and 1847, 3.3 times per 50 years
and spanning six consecutive 50-year cells. The verb vent- occured half as often as
blood (11 cases), but this keyword is tied with blood for the longest span, with six
consecutive 50-year cells, an average of 1.83 occurrences per cell. The noun spleen
occurred nine times, an average rate of 3.0 occurrences across three 50-year cells
in a span of 136 years. Finally, the verb boil- was the most idiosyncratic, with a
total of nine cases: zero cases in the 16th century, two in the 17th, and zero be-
tween 1700 and 1849. Then, boil- reappeared in the 18501899 period and was the
only keyword with cases in the 20th century (four total), with a total of seven
cases between 1850 and 1990, 77.7% of the total cases for the keyword during the
490-year period under study. During that later 140-year period, the longest con-
tinuous span of use for boil- was three cells (18501990), the shortest span of the
four keywords, with an average of 2.33 occurrences per cell.
Metaphor across Time and Conceptual Space

Implications of the frequency results

These results point to two important issues. First, all of the keywords went out of
use before 1850, indicating that some factor(s) influenced the discontinuance in
use. Since we know that the unified model began to go out of use among lay con-
sumers in the 19th century, especially after the publication of Virchows book in
1858, it is possible that the waning use of the cultural model contributed to the
decreasing use of the keywords for metaphors of emotion. Second, the reintroduc-
tion of boil- after 1850, during the period when the unified model was going out of
use, indicates that a different cultural model began to replace the unified model to
provide perspective on the experiential scene for expressing anger. In the dis-
course analysis that follows (see later in this chapter), we will look for indications
of this new cultural model.

Comparing the raw frequencies to the scientific advances

A comparison of the raw frequencies and the normalized rates with the dates of
the scientific advances in human physiology research shows that the use of meta-
phoric expressions for anger varies in concert with some of the scientific advances.
For example, boil- drops out after 1696, 68 years after Harveys 1628 discovery of
the circulation of the blood. Spleen drops out of use after 1736, 108 years after
Harvey and 25 years before Morgagnis 1761 book on the localized origins of dis-
ease within bodily tissue. Blood and vent- drop out after 1847 and 1854, respec-
tively, within 11 years of Virchows 1858 book demonstrating that cell tissue was
the locus of disease. A causal relationship cannot be shown with the available data,
yet the proximity of the drop in use of the keywords to the time in which a major
scientific advance is made known indicates that a correlation between keyword
use and scientific advance is plausible. However, the reemergence of boil- in the
18501899 period, 250 years after it first dropped out and reappearing in the same
50-year cell in which Virchow published his book on cell pathology, contradicts
the historical scientific advances of the time. The discourse analysis of specific
cases (see Data Samples and Analysis, below) provided more insight into the
behavior of boil- and the possible reasons for its renewed use.
In sum, the raw frequencies, the NFR, and the year of the last case for each
keyword shows that the use of the keywords coincides with the rise in popularity
of the unified model in the 16th and 17th centuries and also coincides with the
publication of scientific advances that refuted important aspects of the model in
the 18th and 19th centuries. The next section provides a detailed discourse analy-
sis of selected cases across the historical period, to show the structure, meaning,
and use of the keywords in metaphorical expressions.
Chapter 6. The main study of two diachronic metaphors of anger

The discourse analysis

Discourse analysis of selected cases is presented in order of the date of occurrence,

in chronological order, in each 50-year cell shown in Table 3. The purpose of the
procedure was to investigate the specific syntactic structures, meanings, and uses
of the keywords in metaphorical expressions with full contextual data. The contex-
tual data aided the identification of the target domain in each sample and as well
as the meaning of the linguistic metaphor. The results explain some of the idiosyn-
cratic use patterns noted in the frequency results and also delineate changes in the
framing of the keywords over time, indicating changes in the cognitive conceptu-
alization of anger.

The blood and spleen metaphors and prototypicality

Before presenting the results of the discourse analysis, note that the prototypical
forms of the anger metaphors discussed in previous research (Lakoff & Johnson,
1980, 1999; Lakoff & Kvecses, 1987; see Chapter 2 for discussion) either do not
occur in the corpora or employ a meaning that does not map to the target domain
of anger. Thus, He vented his spleen did not occur in either the P-H or ARCHER
over the 490 year study period; His blood boiled was not found in the P-H, but it
occurred twice in ARCHER. One of these two samples, dated 1665, was eliminated
to resolve the overlap between the corpora; however, discourse analysis of the case
was performed. The case does not use the prototypical grammatical structure of
the linguistic metaphor, and the reference is to sexual desire, not anger.
I observed so many excellencies that my blood began to boyl, and my flesh was
all of a flame. For her hair which naturally curled, and was plaited, was of a bright
flaxen, each hair in the sun glittered like a thread of Gold.

Interestingly, here the speakers skin is described as on fire, and the boiling of blood
does not result in any outward show of anger or violence, unlike the anger CM
described by Lakoff & Kvecses (1987). In sum, only one case1 of the 60 cases ana-
lyzed in the study employed the form/meaning pattern of His blood boiled; none
were found for He vented his spleen.
The result denotes the general trend of the collected keyword samples: the
metaphors employed by the keywords take many different syntactic forms, and the
exact meaning of a case must be derived from the specific situational context, as
Cienki (1999), Croft (2003/1993) and other researchers have suggested. In addi-
tion, the result is an indication of the value of corpus data collection and analysis
what is considered (intuitively) the typical form and meaning by present-day

1. This case is analyzed later in this chapter.

Metaphor across Time and Conceptual Space

researchers is not necessarily typical historically, in the same way that a contempo-
rary native speaker may ascribe a stereotypical form and meaning to a historical
use. Thus, the results of the study indicate that the representative, compiled cor-
pora of actual language use were better guides to the typical form and meaning for
historical native speakers than present-day native speakers. By implication, the
non-linguistic background data more accurately described the meaning and use of
the historical linguistic forms than introspective samples.

Analysis of selected historical metaphor samples

In the discourse analysis section below, additional corpora were analyzed when
the P-H and ARCHER did not provide relevant examples.2 The Modern English
Collection of the University of Virginia Electronic Text Center was employed for
the 15001849 period, and the Making of America Collection of 19th century
British and American magazines at Cornell University was employed for the
18501899 period. The corpus of origin is identified with the samples presented in
the discourse analysis.
A.D. 1500 to 1549. One metaphorical use of the keyword blood was found in
the P-H during the 50-year period, in a translation of Vulgate Latin Old Testament
by William Tyndale published in 1530. The case is shown below.3
And what he sayd: What hast thou done? The voyce of thy brothers bloud cryeth
vnto me out of the erth.

anger is instantiated (through a desire for vengeance), and personification is

used as a device to communicate a metaphorical meaning. However, since this is a
direct translation from Latin, which in turn is a translation from Greek, the use of
the metaphor is not clearly indigenous to native English use of the time; therefore,
the case was eliminated. This result is reflected in Table 3, which shows zero uses
of the four keywords in metaphorical English use during the A.D. 15001549 cell.
The lack of native English metaphorical cases in the P-H corpus in this 50-year
period is interesting and requires some additional analysis to understand the pos-
sible reasons. Non-metaphorical instances of the keywords blood and spleen were
found, and these cases refer to the practice of the unified model. One of these
cases, shown below, is from a medical text by Vicary published in 1548.4

2. See Chapter 4 for detailed discussion of this procedure.

3. As with all cases that are not included for analysis in the main study, the Tyndale case is not
4. This case was presented previously in Chapter 4 in the discussion of the historical Four
Humors model.
Chapter 6. The main study of two diachronic metaphors of anger

These are the places of the humors: the blood in the Lyuer, Choler in the chest of
gal, Melancolie to the Splen, Flegme to the Lunges and the Iunctures, the watery
superfluities to the Reynes and the Vesike.

Vicary was a medical doctor who wrote books on medical diagnosis and treat-
ment, primarily for use by physicians and other experts. The book quoted above
was one of the 31 historical sources considered for the ancillary study discussed
previously; the work was eliminated because it was not written for lay medical
practitioners. In the sample of Vicarys book in the P-H corpus, vent- and boil- were
not found, spleen was found three times, and blood totals 27 occurrences. These
cases comprise 30 of the 44 non-metaphorical cases found in the keyword search
of the corpus.
The blood cases in Vicary discussed or referenced heat, but boil- is not em-
ployed in any of these cases; the following sample describing the heart is typical.
...and the cause of this hollowness is this, for to keepe the bloud for his nourishing,
and the ayre to abate and temper the great heate that he is in, the which is kept in
his concauities.

As discussed previously in Chapter 4, in the unified model the heart was the source
of the bodys natural heat; blood, which had the qualities of heat and wetness, was
stored in the heart, accounting for the hearts high level of heat in the model. These
principles explain the reference to heat (heate) in the sample. In addition, Vicarys
work is a treatise on physical illness and its treatment, accounting for the refer-
ences to physiological blood and the spleen. In sum, the P-H cases show numerous
and explicit evidence of the unified model, but metaphorical cases of the keywords
are not present. The lack of vent- and boil- in the P-H indicates that these keywords
may not have been in use in medical works of the time; however, this conclusion
requires more data to confirm its accuracy.
Although the P-H did not have any metaphorical uses of the four keywords in
the 15001549 period, a search of the University of Virginia Electronic Text collec-
tion for the 50-year period resulted in several cases of the metaphorical use of
vent- and boil-. The sample below is from John Calvins Institutes of the Christian
Religion, published in 1536; three total samples were found. In the case shown
below, anger is the target domain.
Since some feeling of shame restrains them from daring to belch forth their blas-
phemies against heaven, that they may give the freer vent to their rage, they pre-
tend to pick a quarrel with us.

In addition, in Niccolo Machiavellis 1531 Discourses on the First Ten Books of Titus
Livius, a total of thirteen examples of metaphorical vent- were found, including the
one shown below.
Metaphor across Time and Conceptual Space

...all the Plebs departing from Rome, all of which (things) alarm only those who
read of them; I say, that every City ought to have their own means with which its
People can give vent to their ambitions, ...

Here, vent- is mapped to ambition, a character trait, rather than an emotion.

In all, 16 samples of vent- from the two authors were found in the Virginia
Electronic Text collection. The cases target anger, ambition, lust, and other
emotions and desires. Vent- instantiated a variety of emotions in these cases, a pat-
tern that was found in other samples of this period (see below) and also in subse-
quent historical periods.
Similarly, boil- metaphors in the Virginia collection included target domains
for a range of emotions; all 10 were found in Calvins Institutes, including the sam-
ple below.
...our conscience can have no rest at all, no peace with God, no confidence or
security, but is continually trembling, fluctuating, boiling, and distracted; dreads,
hates, and shuns the presence of God.

Interestingly, in this case an abstract concept, the conscience, is boiling, which was
not found in Lakoff and Kvecses (1987) data or in the P-H and ARCHER cor-
pora. In addition, the sample targets several emotions, including fear, hatred, and
shame, which is a wider range of emotion than described by Lakoff and Kvecses,
but more common in the diachronic study. This analysis supports Kvecses con-
struct of the scope of metaphor, in which a single source domain can characterize
many distinct target domains (2010b, p. 136).5 The sample indicates a unique
conceptualization, but this mapping is instantiated via the typical conceptualiza-
tion processes noted in previous CF research. While the sample is not prototypi-
cal, it involves typical cognitive processes. As stated previously, diachronic study
adds important details to the study of cognitive conceptualizations.
Overall, vent- and boil- showed the same characteristics: both were present in
the 15001549 period in the Virginia collection, and both were used to target a
range of emotions, including anger, lust, fear, as well as personality traits like ambi-
tion, and the human body is referenced. However, spleen and blood are not refer-
enced in the Virginia E-text samples. The range of references and targets in the
metaphor samples is broad and variable, unlike the limited and generally fixed
range of the typical forms of the metaphors described by present-day synchronic
researchers. The metaphorical uses of the keywords in this period vary widely in
mapping target domains.

5. As discussed in Chapter 1, we will argue later that a domain matrix provides a more accu-
rate description of the conceptual structure of anger and other emotion concepts; see Chapter 9
for details.
Chapter 6. The main study of two diachronic metaphors of anger

At this point, a reasonable question concerns how widespread anger meta-

phors were at this time. The lack of vent- and boil- samples in the P-H corpus and
the relatively small number of anger metaphors in the Virginia E-text database
indicate that the use of the keywords in English for metaphors of anger in the
15001549 time period were restricted and at the same time employed a variety of
other emotion target domains. As discussed in Chapter 5, the P-H is a representa-
tive compilation of historical English use, and so the lack of samples indicates that
anger metaphors were not widespread in English in the 15001549 period, and
when used, were generally limited to scholarly discourse. In addition, since only
two expert authors in the Virginia collection used the metaphors, then they may
have been part of these authors writing styles, rather than a general linguistic
form. These factors may account for the lack of anger metaphors during the pe-
riod, in both the P-H and the Virginia texts.
A.D. 15501599. A total of three metaphor samples were found in this period
in the P-H; all are for blood. Of these, one sample references the unified model
indirectly. Sample 1 is a reference to the calming effect that a loved one has on the
subjects emotional state.
(1) To heare hir name spoken doth euen comfort my blood.
The unified model is referenced implicitly in the sample because blood brought
healthful physical and emotional benefits. The other two samples both employ the
keyword as a metonymic reference to the human body and to death, as shown in
Sample 2.
(2) But it may be lawfull ynough for wicked men, that thursted the blud of all
the senate & all good men, to seeke our wrak, whom they haue seene de-
fend the good & saue the Senate.
The sample targets bloodlust mapped to the source domain of blood, metonymi-
cally referencing death, in the same way that a predatory animal has bloodlust af-
ter a kill, and so continues to kill. Therefore, the hunting frame is the base of the
metaphor, and the profile is the predators kill (i.e., the senate and good men).
In Sample 3, death is also profiled, but the frame is justice, and Blood met-
onymically refers to the dead body of the murdered man.
(3) I beseech you, consider of me; my Blood will ask Vengeance, if I be un-
justly condemnd.
Through personification, the subjects blood cries for justice against his killer, a
common theme in this historical period and the next. Another possible reading
references blood as a metonymic reference to the mans relatives, who will seek
revenge upon his death. justice is the semantic frame in either case. Overall,
Metaphor across Time and Conceptual Space

the three samples in the 15501599 period involve only one of the four key-
words, with references to the desire for vengeance (also entailing anger), death,
the predatory animal attribute of hunting and killing prey, and justice. As seen
in the previous 50-year cell, anger is one of several target domains for the met-
aphor. More significantly, the emotion is an entailment, not the profile of the
semantic frame.
A.D. 16001649. The frequency of the keyword instances continues to in-
crease, with a total of nine cases spread among all four keywords. Vent-, spleen and
boil- appear for the first time in the collected samples. The samples employed a
variety of target domains, and unlike the 15501599 period, emotions were pro-
filed in all of the cases. In Sample 4, from the year 1614, vent- maps to wrath by
way of a simile.
(4) For when the wickednesse of man was so great, and the earth so filled with
crueltie, that it could not stand with the righteousnes of God any longer to
forbeare, wrathfull sentences brake out from him like wine from a vessell
that hath no vent.
The simile uses the non-metaphorical meaning of vent- to compare Gods righ-
teousness to a container under pressure which explodes; the result of the explo-
sion is the verbal expression of wrath. The example is interesting for its use of
the properties of the blood metaphor, including the container image schema
and pressure. This is the first clear example of the CM in the corpora, yet this
case is different from Lakoff and Kvecses (1987) examples: it uses an older
meaning of vent-, via the noun form. In this case, the cask does not have a vent,
leading to a disastrous result. The structure of the expression is a simile employ-
ing like, rather than a metaphorical source-target mapping. The case is an early
example of the blood metaphor, before the full linguistic metaphor developed
with the verb form of vent-.
Other samples in the period also target emotions. As in the previous historical
periods, all three of the spleen samples map to different emotions, all within the
range of unified models view of the spleen. Sample 5 maps anger, 6 to joy, and 7
to vengeance.
(5) The foole, seeing the pitch ball, pulled to haue it off, but could not but with
much paine, in an enuious spleene, smarting ripe runes after him, fals at
fistie cuffes with him; ...
(6) Whereat the World so tickled her spleene that she was agog, clapped her
hands for joy, and saies she was deepely satisfied, and cryed more.
(7) Now, the cause why this Law was first made, was, for that the women there
were so fickle and inconstant, that, vpon any slight occasion of dislike or
spleene, they would poison their husbands.
Chapter 6. The main study of two diachronic metaphors of anger

The conceptualization instantiating the first two samples is intensity of emotion

is intensity of motion.6 The man in Sample 5 has a ball of tar slapped on his
head (as a joke), and this causes intense anger and leads to a fight. The Oxford
English Dictionary Online (hereafter, OED-O) supports this interpretation: an old,
obscure meaning of envious is [f]ull of ill-will; malicious, spiteful, a meaning
which fits the desire for retribution that the man enacts by starting a fight. In ad-
dition, retribution is Step #5 in the anger prototype scenario described by Lakoff
and Kvecses (1987; see Chapter 2 for discussion). For these reasons, Sample 5 was
analyzed as a metaphorical expression for anger. In Sample 6, the womans spleen
is tickled to the point of causing her to laugh and cry. Recall that in the unified
model the spleen was the origin of a variety of emotions, including anger, sadness,
and merriment; excess black bile over an extended time period was thought to lead
to anger and sudden, extreme violence, and ultimately, insanity and acts of suicide.
Considering the various emotions instantiated in the collected metaphoric expres-
sions, the samples in this period thus reflected lay knowledge of the unified model
as it applied to emotional behavior, mapping the spleen as the source domain to
various target domains of emotion.
The CM proposed here, intensity of emotion is intensity of motion, has
features in common with two primary metaphors identified by Grady (1997):
intensity of activity is heat and intensity of emotion is heat. However, the
intensity of motion CM differs from Gradys two primary CM in one important
aspect: Both of Gradys metaphors target intensity and heat, but heat is missing
in (5) and (6). The difference is due to the influence of the unified model, with the
focus on the cold, black bile of the spleen. There is evidence in the collected data
that intensity of motion is a primary CM because intensity of motion is a
characteristic of both blood and spleen metaphors (see Chapter 2), whereas heat
applies only to blood metaphors. The concept that applies across all of the cases in
this 50-year cell is intensity of motion.
As for the blood metaphor, one of the five samples (8, below) continues the
blood vengeance mapping found in Sample 3 in the previous 50-year period; as
before, the frame is justice.
(8) Seeing myself so near my End, for the discharge of my own Conscience, and
freeing myself from your Blood, which else will cry Vengeance against me.
Sample 8 is similar to Sample 3 in mapping as well as the use of the word Blood, a
reference to relatives who may seek vengeance.
The keyword samples also include two cases of the cold blood metaphoric
expression discussed by Lakoff and Kvecses (1987); however, as found in Samples

6. See Chapter 2 for details on intensity.

Metaphor across Time and Conceptual Space

5, 6, and 7, the targets in Samples 9 and 10 are not restricted to anger they vary
over a range of emotions. These features are shown below for comparison.
(9) ...but the King in Mercy spared you. You might think it heavy, if this were
done in cold Blood, to call you to Execution, but it is not so;...
(10) I was also to see y=r= mother whoe it pleasd not to give me a sighte of her,
but it was happines inoughe for me to convers with y=r= sister Drury, who
talkt at a strange rate, but I had temper to heer her and so parted vpon
fayer termes, onely wishing them a happy retourne, hopeing the Bath wa-
ter would coole ther bloods.
Sample 9 refers to the dispassionate state of mind which allows a human to kill or
murder another person as the result of a calculated, premeditated plan, traits that
were associated in the unified model with excess black bile. Sample 10 is from a
personal letter in which the writer discusses problems in relationships with certain
family members. However, some of the context (such as the reasons for the diffi-
culty) appears to be assumed by the writer to be known to the reader, and so is left
unstated. From the contextual data that is available, the writer conceptualizes emo-
tion on a heat scale (Lakoff & Kvecses, 1987); however, in this case the scale
includes both hot and cold dimensions on opposite ends of the scale, with the dis-
passionate state the writer maintains his temper while speaking with Drury on
the cold, calculating end of the scale and anger on the hot, impulsive side. Thus,
the heat scale includes the full range of temperature ranging from high to low,
from hot to cold; we use the term temperature scale to denote this extended
range. Moreover, Sample 10 references the unified model directly by prescribing a
cure for hot anger: taking a bath in cool water.
reason is also present as a scale, conceptualizing the calculating and impul-
sive traits as opposing ends of the scale, which is also consistent with the unified
model. The scale varies in the degree of reason that is employed by a person. In
Sample 9, the calculating end was instantiated, and in Sample 10, both calculation
and impulsiveness were selected. Taking into account all of these features, the CM
of anger during this diachronic period in English can be characterized as similar
to the analysis offered by Lakoff and Kvecses (1987) for synchronic English, but
more complex in its use of scalar dimensions.7
Finally, the single sample for boil- in the period, unlike the cases found in
Virginia Electronic Text cases discussed in the 15001549 period, targets anger
via personification.

7. The scales of temperature and reason found in these samples will reappear in later time
periods; see the discussion later in this chapter.
Chapter 6. The main study of two diachronic metaphors of anger

(11) Then should he see many grete waters like to drowne him, boilinge and
raginge against him as thoughe they wolde swallowe him up, yet he
thought he did overpasse them. And thes dremes and visions he had every
nighte continually for 3 or 4 yers space.
The bolinge and raginge described here is mapped from the powerful movement of
the sea waves, which is the complex CM, intensity of emotion is intensity of
motion, discussed for Samples 5 and 6. In Sample 11, the container is open,
described as the mouth of a hungry animal ready to swallow (i.e., drown) a victim.
The profile is fear, rather than anger. The conceptualization employs many of
the properties of the blood metaphor, yet the meaning of the concept diverges
significantly. intensity is an important entailment of the conceptualization, simi-
lar to the non-prototypical cases of anger analyzed in Chapter 2 (i.e., intense
response over time)
For a complete view of the period, it must be noted that two of the 10 cases that
were deleted from the dataset (due to the chronological overlap of the two corpora)
map anger to heated fluid. The two samples, one for blood and one for boil, were
found in the same passage from 1693 shown below (Note: the sample is not num-
bered because it was not included in the analysis dataset for the main study).
COURTWITT. Whats that you mutter, ha! pull forth thy Gold. <Draws again.>
Lay it before me to appease my fury, my Wrath boils up, my Blood is all on fire,
And Ill consume the Covetous Race of Mortals.

anger is mapped to heat in my Wrath boils up, and also in my Blood is all on fire.
The mapping is the clear and explicit in these cases. Thus, in the 50-year period,
the mapping of heat to wrath was found in two samples. The mapping clearly
exists in the period and is employed, though the cold blood mapping is more prev-
alent in the dataset.
To summarize, the 10 samples from the 16001649 period vary in their use of
the keywords to target the domain of emotion; the samples conceptualized human
envy, joy, and hatred, animal rage, and two different effects of cold blood. In addi-
tion, the concept of blood vengeance continued from the 15501599 period.
The key concepts mapped in this period were the effects of human emotion and
the desire for justice in human social relations. In addition, Sample 11 displayed a
complex use of elements from the target domain of anger to personify the sea as
a voracious animal, similar to the mapping of anger to a dust storm in the Maalej
(2004) study of Tunisian Arabic, which is metaphorical but not embodied. How-
ever, the profile of human fear is the aspect of the case that is related to human
bodily experience. Sample 11 displayed a complex mix of physical experience,
both human and non-human.
Metaphor across Time and Conceptual Space

Most significantly, the period yielded samples that displayed the temperature
scale, extending from hot to cold. We argue that there are three scales in the sam-
ples: temperature, reason, and control (control is discussed in more detail
in the next section); the presence of reason and control correlate with cold and
dry qualities of black bile and a calculating mindset (i.e., the spleen metaphor),
and absence of reason and control corresponds with the hot and wet qualities of
blood and an impulsive mindset (i.e., the blood metaphor). The three scales pro-
vide an integrated model of the relationship between health, both physical and
mental, and personality traits. In addition, the model of embodiment employed in
the samples reflects the views of human health found in the unified model.
Specifically, the features identified in the metaphor samples are in line with the
view of the unified model that characterized the sanguine and the melancholy tem-
peraments. Thus, the temperature scale, in both the unified model and in the
metaphor samples from the period, extends across emotions from anger to sadness,
unlike the synchronic model of the blood metaphor (see Chapter 2), in which the
heat scale extends within the anger conceptualization from hot anger to cold anger.
The difference is a significant point of divergence, and the issue has implications for
later historical periods (see below) and for the current study (sees Chapter 9). In
sum, the metaphor samples from the 16001649 diachronic period conceptualize
specific principles of the unified model not found in synchronic studies. The unified
model was instantiated in the samples during the 50-year historical period.
A.D. 16501699. The 10 samples of the period develop more details in the uni-
fied model, and they also map the temperature scale to additional emotions,
personality traits, and behaviors, such as grief, drunkenness, and revenge. This
section begins with the vent- samples (both from the year 1688) and the conceptu-
alization of grief.
(12) But, however she was forcd to receive this unwelcome news, in all appear-
ance, with unconcern and content; her heart was bursting within, and she
was only happy when she coud get alone, to vent her griefs and moans
with sighs and tears.
(13) He was forced to retire to vent his groans, where he fell down on a carpet,
and lay struggling a long time, and only breathing now and then Oh
There are several characteristics of the melancholic personality in the unified model
displayed in these samples. First, the temperature scale is employed, mapping cold
(or unheated) black bile.8 Second, despite the absence of heat, Sample 12 employs

8. All of the vent- samples in the dataset lack heat, further evidence for the temperature
scale discussed in the previous section.
Chapter 6. The main study of two diachronic metaphors of anger

the word bursting to characterize the feeling of grief, and the object of venting is
moans, sighs, and tears. In the unified model, pressure within the spleen is the result
of increases in the fluid volume; heat and steam are not instantiated, even though the
fluid in this case is warm and wet blood; heat and steam are absent because the target
domain is grief, not anger. Third, in both samples the person who suffers from
melancholy vents in private rather than in public, a significant difference with the
blood metaphor concerning the expression of emotion, which precludes the possi-
bility of retribution, in line with the spleen metaphor. The melancholy person was
typically a lover of solitude, and this principle logically explains the private venting
of emotion.9 Finally, in both samples, the bursting of the container results in non-
verbal expression of grief and sadness, including moaning, sighing, crying, and
breathing (possibly heavy and labored). These characteristics follow the unified
model, and they also differ in these respects from the blood metaphor.
The absence of heat and violent behavior directed outward towards others are
also evident in the two spleen samples, one of which is shown in Sample 14.
(14) We were dull Company at Table, worse A-bed. Whenever we met, we gave
one another the Spleen. And never agreed but once, which was about lying
The sample also follows the basic principle in the unified model of referencing the
spleen to indicate sickness. Through the phrase gave one another the spleen, dislike
(cf. sample 7) is mapped to illness via a metonym.
The blood samples in the period continue the cold blood mapping discussed
in the previous section, and also develop details about the effects of heat on emo-
tion, desire, and behavior. These samples (15, 16, and 17) are shown together.
(15) Come, Lory, lay your Loggerhead to mine, and in cool Blood let us con-
trive his Destruction.
(16) Whereupon although present and privat Execution was in rage done upon
Edric, yet he himself in cool blood scrupld not to make away the Brother
and Childern of Edmund, who had better right to be the Lords Anointed
heer then himself.
(17) And though in cold blood he was a generous and good natured man, yet
he would go far in his heats, after any thing that might turn to a Jest or
matter of Diversion: He said to me, He never improved his Interest at
Court, to do a premeditate Mischief to other persons.

9. These characteristics of private expressions of emotion without seeking retribution are

similar to those found in the Geeraerts, Gevaert, and Speelman (2011) study discussed in
Chapter 3. These characteristics will be found in later time periods; see later in this chapter.
Metaphor across Time and Conceptual Space

The words cool and cold have the same conceptualization in these cases. Sample 15
continues the dispassionate, premeditated disregard for human life that was discussed
in the previous section. Notice also that the preposition in is used in all three sam-
ples, a reference to both the container and to the variable nature of physical and
mental health on the temperature and reason scales. Sample 16 conceptualizes
the quality of reason entailed in cool blood, which preserves positive regard for
others (by the omission of a violent act of murder). Finally, Sample 17 takes the
reason quality and extends it to the commission of generosity toward and good na-
tured interaction with people, an extension of reason into good deeds. The pro-
gression through these concepts employs the scales of temperature, control, and
reason discussed in the previous section and builds new entailments within them.
These results support the conclusions of Sim (2011)10 regarding the negative
and positive connotations of in cold blood. Two of the three instances found instanti-
ated positive connotations, which is a higher rate of occurrence than Sim found in
present-day American English data. The positive connotation may have been more
common in historical English, but more samples are needed to confirm the result.
Two other blood samples are interesting for adding details to the tempera-
ture scale this time, on the hot end of the scale.
(18) And the natural heat of his fancy, being inflamed by Wine, made him so
extravagantly pleasant, that many to be more diverted by that humor,
studied to engage him deeper and deeper in Intemperance: which at length
did so entirely subdue him; that, as he told me, for five years together he
was continually Drunk: not all the while under the visible effect of it, but
his blood was so inflamed, that he was not in all that time cool enough to
be perfectly Master of himself.
(19) To this he answered, A man could not write with life, unless he were heat-
ed by Revenge: For to make a Satyre without Resentments, upon the cold
Notions of Phylosophy, was as if a man would in cold blood, cut mens
throats who had never offended him:...
Sample 18 is a clear reference to the unified model in its use of the word humor.
Different foods were believed to have the hot/cold and wet/dry qualities; Boordes
(1542) book Dyetary of health and other historical sources for the composite
model contain information on the qualities that many foods and beverages were
believed to possess. Wine was viewed as a hot drink, increasing the heat of
the blood and affecting physical health and behavior. The sample shows these re-
sulting behaviors: increased blood heat, a warm pleasantness, and a lack of emo-
tional control due to decreased reason. The sample employs the metaphorical

10. See Chapter 3 for a review of this study.

Chapter 6. The main study of two diachronic metaphors of anger

extension burn and applies it to blood, though in the unified model blood could
not burn because the fluid had the quality of wetness.11 Sample 18 is consistent in
its use of the three scales found in earlier 50-year periods.
The control concept is an important mapping for the temperature scale in
the unified model because increased heat led to increased anger and physical vio-
lence. The concept is a scale, ranging from no control (impulsiveness) on the hot
end of the scale to total control (calculating) on the cold side. Therefore, the
control scale corresponds to the temperature and reason scales discussed
The control mapping found in the samples parallels the force/control
scale found in the Koivisto-Alanko and Tissari (2006) study, discussed in Chapter 3.
In that study, reason was associated with the control side of the scale, and
emotion was associated with force. In the current study, similar associations
were found, and the temperature scale adding further details: heat is associated
with intensity (of motion, a type of force), and cold with control. However,
while support was found for Koivisto-Alanko and Tissaris results on the reason
scale (i.e., reason is cold, unreasonable is hot), in the current study emotion was
found on both ends of the temperature scale, rather than on the heat end only.
Sample 19 follows the same conceptualization by mapping heat to revenge,
a desire which can lead to physical violence. In addition, revenge against an of-
fender is viewed more positively than violence done in cold blood, where the vic-
tim is usually not the offender. Retribution has positive qualities that reside only
on the heat side of the temperature scale.
Finally, the single sample for boil- in the period shows an important aspect of
heated fluid.
(20) Sir Tun. Oh, Ill warrant you my Hero, young Men are hot I know, but they
dont boyl over at that rate, neither;...
Boiling over is the result of heat, yet the use of the preposition over indicates that
the container is open, like an uncovered pot. The feature is different from the
closed container found in synchronic studies of the prototypical CM of anger, but
it does fit the cooking extension found by Lakoff and Kovecses (1987),12 which
they deemed non-prototypical. This particular extension, referred to here as the
cooking semantic frame, was also found in six of seven samples for boil- between
1850 and 1990. The semantic frame will be discussed again in more detail in that
section of this chapter.

11. The quality of wetness and its effect on emotion will be discussed in greater detail later in
this chapter.
12. See Chapter 2 for discussion of the cooking extension.
Metaphor across Time and Conceptual Space

To summarize the samples of the 16501699 period, the temperature scale

continues to span across a variety of emotions, from anger to grief. In addition, the
control scale was further developed. control correlated with the tempera-
ture and reason scales, ranging from no control and unreasonable on the
heat side to total control and reasonable on the cold side of the continu-
um. Additional entailments were found for heat, including the hot/cold and wet/
dry qualities of foods (e.g., wine) in the unified model, and the mapping of heat
with revenge as a cause of violent retribution for injustice. Hot revenge was also
contrasted with cold-blooded murder, with revenge given the positive evaluation
because righting a wrong against an offender is justified, whereas killing an inno-
cent victim is not; conversely, reason and control were mapped to positive per-
sonality traits, including generosity and good humor. Emotions on both sides of
the temperature scale entailed both positive and negative connotations, and this
is consistent with the unified model. Finally, the cooking semantic frame was
employed for the boil- sample found in this period; the conceptualization will be
seen for the boil- keyword in other 50-year cells, beginning with A.D. 18501899.
A.D. 17001749. The data samples now shift to the ARCHER corpus for the
17001990 period. In the 17001749 cell, seven samples were collected for vent-,
spleen, and blood; boil- dropped out at this point in time.13 The first of the samples
shown is for vent-.
(21) This confirmation of what Liberius had said and the jeers he had put upon
him were such a weight upon the haughty spirit of Theophilus, who had
the exact temper of some fellows of colleges, that it made him very chagrin
and full of spleen, insomuch that he was obliged to retire to his chamber
where he vented these expressions: Who could have divined that Sylvia
was a gentlewoman? Tis seldom persons of fashion turn beggars...
This is the first sample that clearly displays the venting of verbal expressions, as
opposed to non-verbal tears or sighs, and it includes the lexical item spleen, as well.
Again, as found in the previous periods, heat is not associated with venting spleen;
the emotional expression centers on sadness (i.e., chagrin) instead of anger, and
expressing the emotion is expressed quietly in solitude, compared to the violent
and public expression of anger. The ARCHER sample is consistent with both the
unified model for the spleen and with the samples in the PPCEME corpus for the
previous 50-year cells.
Another sample, 22, again shows the mapping to sadness.

13. As noted previously, boil- returns in the 18501899 period. See that section of this chapter
for discussion.
Chapter 6. The main study of two diachronic metaphors of anger

(22) As soon as you had reached the house, I shifted my material figure for one
more becoming the dignity of the celestial condition; and being again in-
visible, I heard the fantastic relation you gave your brother, who told you,
twas all the effect of the spleen and obstinate grief you had indulged since
my death:...
In contrast to 22, Sample 23 maps the spleen to the concept of a long-term com-
plaint or grudge against another person.
(23) An old Spleen she had a long time bore to Yamatalallabec, on account of
his Friendship with a Person at enmity with her, tho he had never assisted
him in any Designs against her, made her gladly enter into the Measure
Oudescar had taken for the Establishment of his Favourite:...
The concept of grudge is a common one in spleen metaphor, and one that the mel-
ancholy temperament was believed to possess. envy and jealousy are also
mapped in spleen metaphors, both of which are logical causes of grudges. A cold
private grudge contrasts with the hot public revenge discussed in the previous
section temperature is on the cold (unheated) side of the scale, so violent be-
havior is replaced by private scheming (Designs).14 In addition, the time scale
for retribution is longer than for revenge, due to the control scale. The scheming
person waits for an appropriate opportunity to fulfill the grudge rather than take
immediate retribution as the prototypical anger scenario suggests. Based on these
characteristics, the grudge metaphor follows the controlled intensity over
time CM proposed in Chapter 2 for the non-prototypical cases of anger.
The blood samples of the 17001749 period continue to entail cold blood
and its effects on emotion, reason, and control on behavior. Sample 24 is an
(24) GAY-LOVE. Then turn back and use your Sword for now my Blood is
cool, Id rather lose <loose> my Life than lose <loose> your Friendship.
BELLMOUR. I cannot look on thee, and bear resentment; Ill never meet
thee more but thus <(embracing him)> this is real and all my Angers
As was seen in previous samples, the entailments of reason and control are
implicated in the sample; Gay-Love no longer wants to fight his friend because his
hot-blooded anger has moved to the cold end of the temperature scale, bringing
emotional control via reason.
The next sample profiles imagination in the melancholy temperament.

14. Sample (23) is similar in these characteristics to (13), and both support the analysis of
Geeraerts, Gevaert, and Speelman (2011) concerning the private and non-violent form of anger.
Metaphor across Time and Conceptual Space

(25) But though I could easily argue these Sir Gravities down, though a sen-
tence or two would do their business, put them beyond the power of re-
plying, and strike them dumb, yet do I think it not worth my while; their
greatest and most wonted objection against these Eudemons and Kakode-
mons, being, that it arises all from the work of fancy, in persons of a mel-
ancholic blood.
Imagination (i.e., the work of fancy) and intelligence were specific traits of the mel-
ancholy temperament. Burton (1932/1621) states that writers, clergy, and scholars
were believed to be melancholy professions due to the association of these voca-
tions with intelligence, imagination, and solitary work. Again, these associations
support the entailment of reason at the cold end of the temperature scale.
Summarizing the results of the 17001749 historical period, the details of the
mapping of vent- and spleen to sadness were delineated in more detail. In addi-
tion, the concept of cold grudge, in contrast to the heat inherent in revenge in
the previous period, was established. Finally, the concept of cold blood, with the
entailment of reason, continues to be a common theme in the study dataset.
A.D. 17501799. The number of collected samples dropped in this period to a
total of seven, and spleen dropped out of the dataset. Samples were found for two
keywords vent- and blood (as mentioned previously, boil- dropped out in the
17001749 period). For the first time in the dataset, vent- is mapped to violent,
public anger, as shown in Sample 26, published in 1788.
(26) By the duchesss earnest solicitude to please, she destroyed her own pur-
pose, and her obedience, like water flung upon a raging fire, only inflamed
her husbands follies; and therefore, when he was in an ill humour, the
duke vented his rage on her. He did not care how often he quarrelled with,
or, to speak more properly, how often he insulted her;...
In contrast to previous samples of employing the lexical item vent-, the anger is
expressed publicly. However, consistent with spleen metaphor samples from the
previous periods, verbal expressions were vented (e.g., insulted her), rather than
physical behaviors associated with hot anger, such as skin redness or bodily agi-
tation. In addition, recall that in the 16001649 period, vent in the noun form was
used to describe the wrath of God within a closed container that is about to burst
from the pressure (see Sample 4). In comparison, Sample 26 is the first in the data-
set to ascribe the venting of anger to a human being, and to do so with the verb
vent- in the modern verb-object syntactic structure. Moreover, the unified model
is also referenced in the words ill humor. The sample is consistent with the prin-
ciples of the unified model in characterizing anger via the spleen metaphor.
As Sample 26 indicated, anger was possible in the melancholic temperament,
though it was characterized differently in comparison to the hot and wet anger of
Chapter 6. The main study of two diachronic metaphors of anger

the sanguine temperament. Melancholic anger was called melancholy adust, a

medical condition in which cold, dry black bile was heated to the point of being
burned, and the extreme heat resulted in extremely violent behavior and insanity.
Suicide and violent crimes were thought to result from melancholic anger. It is
unclear from Sample 26 if the Duke was thought to be suffering from melancholy
adust, but the reader of the time would have understood that such a result was
possible after prolonged exposure to the effects of the condition.
A sample for blood, published in 1786, extends the mapping of burning to
(27) ALEXIS. Oh you traitress artful slut! this must be all a feint. I clearly
heard she feels it too, that she must concern my wife, or my daughter oh
my blood burns! She feels it too!
Similar to Sample 26, Sample 27 maps burning fluid to anger or sexual desire; the
speaker displays angry behavior of cursing but states She feels it too!, possibly a
reference to Alexis own feelings of love for the traitress in either case, the emo-
tions instantiated are clearly intense and impulsive. The sample extends the burn-
ing property of black bile to blood. The extension violates the tenets of the unified
model (recall that blood had the quality of wetness and so could not burn), yet the
extension is used in several of the collected samples discussed previously. Sample
27 is included in that group.
Though Samples 26 and 27 are the first examples of vent- to indicate anger in
the analysis dataset, these were not the first found in the compiled corpora. One of
the samples, published in 1724, eliminated to resolve the overlap in chronology
between the corpora, has the same mapping.15
And indeed mens spirits were so sharpened upon it, that we all looked on it as a
very great happiness that the people did not vent their fury upon the papists about
the town.

The sample shows the mapping of venting to anger, though no explicit connec-
tion to melancholy adust is made. Overall, Samples 26, 27, and the unnumbered
sample display vent- to indicate anger, and Sample 26 is consistent with melan-
choly adust, a medical condition of the spleen and black bile in the unified model.
Finally, two samples for blood from the 1790s show the range of the tempera-
ture scale discussed in previous periods.
(28) In cool blood, yet with firm attachment, we now see blended in her, the
peerlessness of enterprise, the deportment, ardor and heroism of the vet-
eran, with the milder graces, vigor and bloom of her secreted, softer sex.

15. The sample is not numbered because it is not included in the study dataset.
Metaphor across Time and Conceptual Space

(29) Valmont, whose imagination, long fixed to one point, had seen nothing in
her confinement but a plan to deprive her of some envied advantage of
rank or fortune, now gazed, as her blushes and tremor heightened her
beauty, with a consciousness of it he had not before felt; and no sooner did
his mind catch a ray of truth, than it became perfectly enlightened. All the
warm blood congealed round his heart, flowed obedient to the voice of
humanity; and in the wild hope of affording protection, he seemed to have
forgotten how much he wanted it.
Sample 28 is consistent with previous samples concerning the mapping of cool
blood to reason. The speaker is able to see the womans positive attributes when
the speaker is in the cool condition. Sample 29 shows the effects of warm blood
(as opposed to hot blood) on the feeling of positive fondness towards another
person. These two samples show, similar to previous samples, that the tempera-
ture scale ranges across emotion categories; the scale is not confined to anger.
In sum, the 17501799 period provides examples which map the action of
venting to anger, including the medical condition melancholy adust. Also, the
blood samples provide more details about the temperature scale in the unified
model, consistent with samples from previous periods. Overall, through the end of
the 18th century, metaphors employing the keywords consistently and systemati-
cally entail fundamental principles from the composite model found in the 18
historical source texts.
A.D. 18001849. There are five metaphor samples in the first half of the 19th
century; one for vent- and four for blood. The vent- sample (shown below),
similar to the cases discussed for the 18th century, specifies a verbal expression
of emotion, as opposed to a non-verbal physical act, consistent with the spleen
(30) I came only to sell a few apples, said Mary. Heaven has sent that girl to the
rescue of my life, said Butler, under the impulse of a feeling which he could
not refrain from giving vent to in words.
Several samples in the dataset specifically identify the use of words when venting
emotion, apparently as a means to separate the verbal from the non-verbal. Yet, for
blood and boil-, keywords for the blood metaphor, an identification of the mode of
expression was not found for any sample in the dataset. The consistency of the vent-
samples on this point indicates the writers intent to identify the type of anger that
fits the context of the discourse situation. Metaphors with blood and boil- entail a
non-verbal act of violent retribution, whereas spleen and vent- metaphors entail a
verbal act expressing an emotion or a non-violent grudge. These characteristics are
consistent with the unified model, the blood and spleen metaphors proposed in
Chapter 6. The main study of two diachronic metaphors of anger

Chapter 2, and the results of the Geeraerts, Gevaert, and Speelman (2011) study
discussed in Chapter 3, for these two prototypes of anger.
The blood samples for the period continue to specify aspects of the tempera-
ture scale. Sample 31 demonstrates the effects of cold blood on emotion.
(31) It was evening when I reached the hills of Languedoc, and looked impa-
tiently toward my cheerful home beneath. I looked the last sun-beam
glared redly upon smoking ruins! Oh! oh! the blood now chills and cur-
dles round my heart the wolves of war had rushed by night upon my
slumbering fold fire and sword had desolated all. I called upon my wife
and my infant. I trampled on their ashes while I called!
In the unified model, the emotion of fear chills the blood and causes the blood to
rush from the skin causing the skin to turn pale to the heart, in turn causing
the heart to feel cold; the process is described in the words the blood now chills and
curdles round my heart. The cold side of the temperature scale extends to the
emotion of fear.
The next sample entails aspects of hot blood.
(32) When months ago you slept under my roof ay, slept what should have
hindered me from stabbing you during the slumber? Two nights since, when
my blood was up and the fury upon me, what should have prevented me
tightening the grasp that you so resent, and laying you breathless at my feet?
Sample 32 uses the phrase my blood was up to entail the unified model principle
that blood rushed toward the skin and head during an expression of anger, which
is the opposite direction of the flow of blood when expressing fear in Sample 31.
The concept can be termed anger is up, related to the blood metaphor; the latter
underlies Lakoff and Kvecses (1987) analysis of the CM of anger. anger is up
was also found in one of the eliminated samples, my wrath boils up.16
To summarize the 18001849 period, the samples entailed the unified model
principles that cold blood is produced by fearful life situations, causing physical
symptoms such as pale skin and cooler body temperature; and, that fear and
anger cause the blood to flow inward toward the heart and outward toward the
head and skin, respectively.
A.D. 18501899. Boil- reappeared in this period after an absence of 150 years
with three samples, vent- had two samples, and blood dropped out after a span of
almost 300 years. In addition, continuing a trend from the previous period, both
of the vent- samples mapped to the target domain of anger. Sample 33, below,
specifies wrath.

16. Discussed after sample 11 in the section of this chapter concerning the 16501699 period.
Metaphor across Time and Conceptual Space

(33) Will, we are compelled to say, did not really care a copper for Donsy, and
he bore no real ill will to Lanky: but when he found himself thus ignomi-
nously [sic] abandoned, his authority despised, his rival preferred, he fell
into a passion and looked around him for some means of venting his
Sample 34, unlike previous boil- samples, does not map to emotion but to ener-
gy, a new target domain in the dataset. The passage was published in 1872.
(34) Bunny was heavy and sleepy therein, and did nothing but yawn and stretch
out her arms. Barbie, on the other hand, was ready to boil over with de-
light and liveliness, flashing about like a little dab-chick.
This case is similar in some respects to Sample 11, where the CM was identified as
intensity of emotion is intensity of motion. In Sample 34, the CM is inten-
sity of energy is intensity of motion (an extension of the CM in Sample 11
with the target a positive attribute of personality), whereas the target in Sample 11
is a negative emotion. The base of the frame is movement (the extension of mo-
tion across a physical location), and the profile is energy required for move-
ment. Additionally, Sample 34 employs the lexical phrase boil over, entailing a
container open at the top, also found in Sample 20.17 Both samples instantiate
the cooking semantic frame.
Two other boil- samples from 1889, Samples 35 and 36, add new extensions,
also from the cooking semantic frame.
Yet stay before I enter into particulars,
allow me to give you an insight into the state of my mind,
- Mr. Go tightly!
Go-lightly, sir, I never do go tightly!
You see before you a man, furious with
indignation, sir, literally boiling over!
Well, sir, Id advise you to wait till you simmer
down a little. Its as well to appear cool and
collected before people
but, I confess, I wouldnt have his wife show her face
at this moment, for a very considerable trifle!

17. See Sample (20) for further discussion of the open container.
Chapter 6. The main study of two diachronic metaphors of anger


I see you are dying with curiosity to know what has
excited my anger, which I consider both inquisitive and
My dear sir, you are mistaken; I dont care one straw
about you or your anger either. You may boil all away,
as far as Im concerned.
As with previous samples of hot anger, the temperature scale extends from un-
reasonable anger to reasonable calmness (i.e., cool and collected), but both of the
samples use terms associated with boiling food in the cooking semantic frame:
boiling over, simmer down, and boil all away. The cooking frame was found previ-
ously in Sample 20 from the 16501699 period; in that case, the frame was applied
to the young men as cook pots that may boil over. Also, in Sample 34 in the current
50-year period, the frame was applied to a woman. In Samples 35 and 36, the same
frame and words are applied to an adult man. The cooking semantic frame is
therefore not a new innovation in the 18501899 period, but additional lexical
items from the cooking semantic frame are employed in the cases for this 50-year
Recall from Chapter 2 that Lakoff and Kvecses (1987) specifically comment-
ed about the use of words like simmer in the anger CM, arguing that the cook-
ing entailment is a minor, non-prototypical one. However, from a diachronic
point of view, they do not appear minor or atypical since the cooking frame was
employed in a total four different samples spanning more than two hundred years
(the cells from A.D. 1650 to 1899). In addition, one of the eliminated ARCHER
samples, from 1692, also used boil over:
I could never get any Body to give me a satisfactory Reason, for her suddain and
dextrous Change of Opinion just at that stop, which made me conclude she could
not help it; and that Nature boild over in her at that time when it had so fair an
opportunity to show itself...

The sample does not instantiate anger, but cooking is clearly present. It is also
worthy of note that the fact that a total of seven samples out of the total of 50 in
the study dataset employed cooking (three more will be presented in the
19001949 and 19501990 periods). Considering the use of the concept in sam-
ples in the dataset and the variety of lexical items that instantiate it, we argue that
cooking is not a minor entailment but an important semantic frame that moti-
vates a significant number of cognitive concepts, including anger in the blood
and spleen metaphors
Metaphor across Time and Conceptual Space

The examples from the 18501899 period showed some important new exten-
sions which previous historical periods did not exhibit, by including the target
domain of energy, new to the conceptualization of anger, and employing the se-
mantic frame of cooking with more frequency and a wider variety of lexical items.
A.D. 19001949. One sample of the boil- keyword was found in this period,
and one for blood + boil-. Vent- dropped out in this period, after being in continu-
ous use for over 250 years.
The boil- Sample 37 was similar in conceptualization to Sample 34, intensity
of energy is intensity of motion. The passage was published in 1931.
(37) As for Ethel Smyth, whom you must meet, she has boiled over with a kind
of effervescence of force playing the trombone, golf, conducting, walk-
ing, riding, singing, loving, all at the same moment, so that she has, or had,
a temperature of 104 and is nursed by a single maid with Lady Betty at
the bedside.
This is the same conceptualization found in Sample 34; effervescence historically
was related to boiling, according to the OED-O, but the resulting behaviors en-
tailed (e.g., playing trombone, golf, and loving) are related to energy and
excitement rather than to anger. The fact that the same CM, intensity of en-
ergy is intensity of motion, occurred in Sample 37 indicates that the concept
was not a one-time, novel creation in Sample 34. Sample 37 was also the fifth ex-
ample of boil over found in the corpora, employing the cooking semantic frame.
The blood + boil- sample, (38), was the only one found in the dataset that em-
ploys the prototypical syntactic structure and lexical items of the blood metaphor.18
(38) Dont call me `sir. Call me Comrade. Do you know what you are, my lad?
Youre an obsolete relic of an exploded feudal system.
Very good, sir.
If theres one thing that makes my blood boil in my veins
Have another sardine, chipped in young Bingo...
The fact that Sample 37, a relatively new conceptualization, appears in the same his-
torical period as the typical concept in Sample 38, demonstrates the variable nature
of conceptualization. The 20th century was characterized by variations in the con-
ceptualization of emotion which recall old cultural views and also employ new
A.D. 19501990. Again, as in the previous period, only boil- samples were
found. Two of the three samples employ terms from the cooking semantic frame,
including Sample 39, a newspaper report published in 1967.

18. As discussed previously, a sample from 1693 which was eliminated to resolve the chrono-
logical overlap between the two corpora also was of the blood + boil- type.
Chapter 6. The main study of two diachronic metaphors of anger

(39) AFTER MONTHS of intensive struggle between factions of the Chinese

Communist Party, the unrest in the country has boiled over into some-
thing approaching civil war. A series of reports from Peking yesterday
spoke of violent clashes between party groups resulting in a death roll of
more than 50, ...
In this sample, boil over is the same verb used in Samples 34, 35, and 37; in Sample
36, boil away is used instead. All of these samples instantiate the cooking frame.
In Sample 39, the cooking frame is extended to a large group of people (without
a container), employing the intensity of emotion is intensity of motion
CM also found in Samples 34 and 37.
Sample 40 below also activates the cooking frame, from a novel published in
(40) (To DR. PRENTICE.)
If this boy becomes foul-mouthed keep him on the boil till
I return. (Goes to the garden, followed by MRS. PRENTICE.)
Again, keep him on the boil entails a cooking pot in the cooking frame for an
emotion that is not stated explicitly (though it is probably anger).
The final period in the discourse analysis, 19501990, continues the trend to-
ward the enactment of the cooking frame in metaphoric expressions of anger.
The profile in both cases is an open container or cooking pot, sometimes refer-
ring to the human body (see Samples 34, 35, 36, 37, and 40), and other times ref-
erencing a group of people, as found in Sample 39. Finally, the CM, intensity of
emotion is intensity of motion was found to motivate Sample 39; the same
metaphor has been found in other historical periods.
The last 140 years of the period under study revealed several important fea-
tures of the conceptualization of anger. First, the repeated characterization of the
human body as an open container diverges from the prototypical CM of anger
in which the human body is a closed container under pressure. The cooking
semantic frame thus does not allow for the container to explode; instead, the fluid
in the container boils over or boils away to relieve pressure on the fluid. In
addition, the samples conceptualized several different emotions, including anger
and excitement. Finally, six samples out of eight (75%) during the period em-
ployed the cooking frame, and overall seven out of the total nine boil- samples
(77.7%) in the total study dataset employed this frame. An NFR of 3.16 for the six
cooking frame samples in the ARCHER corpus is evidence that this frame is
typical, not atypical, in historical English use since 1700. The frame is also an old
one; the OED-O lists uses of boil- that employ the cooking frame as early as the
14th century A.D.
Metaphor across Time and Conceptual Space

Chapter summary

In each 50-year period analyzed, the metaphor samples display a wide range of
emotions, grammatical structures, and situations of use. The form/meaning pairs
found in the collected samples are different in several important ways from the
blood metaphor proposed by Lakoff and Kvecses (1987) and other studies. The
major results are summarized below.
1. The keyword searches of the two text corpora resulted in the collection of 62
linguistic metaphors of anger; 12 were eliminated for various reasons, leaving
a total of 50 samples for the analysis.
2. The frequency analysis (see Table 3) of 50-year periods indicated that a) the
first instances of the metaphors in the texts occurred in the 15501599 period
and ended in the 19501990 period; b) the four keywords were variable in
their occurrence over time, but all initially dropped out before the 18501899
period. The Normalized Frequency Rate (NFR) analysis showed that the oc-
currences of the four keywords increased in every 50-year cell from 1550 to
1699 and then decreased in every cell after 1700 until the 18501899 period;
only the verb boil- increased in occurrences at any time after 1700. The key-
word boil- began to occur for a second time in the 18501899 period; the final
instance was found in the 19501990 period.
3. The frequency results indicated that, after 1850, a different cultural model re-
placed the unified model to provide perspective on the experiential scene. This
conclusion also coincides with the results of the ancillary study concerning the
influence of the scientific advances in human physiology, especially Rudolph
Virchows book in the 1850s (see #5 below).
4. The CADS analysis of all of the cases found that the linguistic metaphors em-
ployed specific aspects of the unified model of human health, the blood meta-
phor, and the spleen metaphor from 1500 until 1850. After 1850, other cultural
models provided perspective on the experiential scene.
5. The CADS analysis of the 18501990 samples (all for boil-) showed that the
conceptual structure of six of the eight collected cases employed the cooking
semantic frame, which has several features that are different from both the
unified model and the blood and spleen metaphors. This historical shift to a
different cultural model and the resulting change in perspective on the expe-
riential scene coincides with the decreasing popularity of the unified model
over time, as shown in both the ancillary study of non-linguistic data
(see Chapter 5) and the decreasing frequency of use of the blood and spleen
metaphors after 1700 (see Table 3). The data indicate that the cooking seman-
tic frame replaced the unified model in the mid- to late 1800s.
Chapter 6. The main study of two diachronic metaphors of anger

6. The prototypical forms of the blood and spleen metaphors (as described in
synchronic studies) were rarely employed in the diachronic data. One instance
of the prototypical structure of the blood metaphor (i.e., His blood boiled) was
found during the 490-year study period. That case occurred in the 19001950
period, further indicating that the blood metaphor is of recent origin and/or is
motivated by a contemporary conceptualization of anger. Prototypical in-
stances of the spleen metaphor were not found in the dataset.
7. These results indicate that the blood and spleen metaphors were prototypical
forms of anger during the historical period under study, but that prototypical-
ity is a characteristic of a specific cognitive conceptualization, not the linguis-
tic form employed to express the concept. As previous studies have shown,
linguistic form is highly variable over time and so is not a reliable indicator of
prototypicality. This conclusion supports the results of many cognitive-func-
tional studies that metaphor is a construct of the human mind, rather than a
language form (see Lakoff, 1993).19
8. The changes in conceptualization and the cultural model may be attributed
to changes over time in the unified model and its decreasing popularity
among lay people, due to scientific advances in human physiology during the
490-year study period.
Before discussing the implications of the above results for the main study of his-
torical metaphor, Chapters 7 and 8 describe additional micro-studies of concep-
tualization and cultural models, focused on historical English data from the 19th
century. The results of these studies support the main study and also extend the
results of the main study to other theoretical issues currently discussed in CF re-
search. Following the micro-studies, Chapter 9 presents the answers to the research
questions for the main study of historical metaphor; the implications for CF theo-
ry and metaphor research and recommendations for future research are also dis-
cussed. A brief justification for applying the results of the studies in this volume to
second language teaching pedagogy is presented in the Epilogue.

19. However, the results also indicate several additional points that contradict current theory
and research practice in conceptual metaphor. These points are discussed in Chapter 9.
part iii

Micro-studies of emotion the 19th century

chapter 7

The edge of anger

The spleen metaphor across emotion domains


As shown in Chapter 2, Lakoff and Kvecses (1987) analysis of the CM of anger

places spleen metaphors on the periphery of the cognitive domain of anger in
English as a minor, non-prototypical form. Yet, the main study of historical meta-
phor (see Chapter 6) found significant differences between the spleen metaphor
and other metaphors of anger that warrant further study. The purpose of this
chapter is to describe a historical study of the spleen metaphor. The following sec-
tion describes the purposes of the study.


This study has three purposes. First, the study of a CM on the periphery of a cogni-
tive category can provide useful information. Geeraerts (2006, p. 156) states that
[f]rom a methodological point of view, the periphery of natural, non-uniquely
definable categories is as interesting as their salient center(s), because it is pre-
cisely the relationship between both that typically characterizes natural categories.
Cognitive Linguistics is not only interested in what constitutes the centre of a cat-
egory, but also in how this centre can be extended towards peripheral cases, and
how far this extension can go.

Second, historical study can inform knowledge of present-day language use. As

discussed in Chapter 3, cognitive linguists have noted the intimate relationship
between diachronic and synchronic linguistic forms (e.g., Allan, 2008; Bybee,
1988, 2001, 2003; Bybee, Perkins, and Pagliuca, 1994; Kvecses, 2000; Sweetser,
1990). Historical study of spleen metaphor may thus reveal aspects of present-day
conceptual metaphor which synchronic study may not show. Such study can con-
tribute to conceptual metaphor theory, the process of cognitive conceptualization,
and the role of cultural models.
Finally, a close study of the spleen metaphor may reveal details of the concep-
tualization that either support or refute the findings of the main study of historical
Metaphor across Time and Conceptual Space

metaphors of anger. This study of the spleen metaphor was conducted in order to
investigate the validity of the main study results. Overall, the purpose of the cur-
rent investigation is to explore the spleen metaphor in the conceptual domain of
anger in English, in order to further understand the relationship between cogni-
tive conceptualization and cultural models and the specific content of the spleen
Two research questions will be investigated:
1. What are the specific features of spleen metaphor conceptualization?
2. Is the conceptualization an elaboration of the synchronic blood metaphor or a
separate conceptualization?


The study employed natural language data from English native speaker historical
corpora, analyzed via discourse analysis. Studies of conceptual metaphor using
natural language data, collected from compiled corpora, both general (e.g., Deig-
nan, 2003; Deignan & Potter 2004; Charteris-Black, 2003), and specialized (e.g.,
Moder, 2004, 2008; Caballero, 2003; Stefanowitsch, 2006b) have shown promise
for delineating how conceptual metaphor is instantiated by language users. Infor-
mation on language use can increase understanding of the conceptualizations that
inform the linguistic expressions.

Data collection

The study data were collected from two Internet digital corpora of historical British
and American magazines. One site is the Internet Library of Early Journals
(hereafter, ILEJ), a cooperative project initiated by the universities of Birmingham,
Leeds, Manchester, and Oxford to digitally preserve early British magazines from
the mid-eighteenth century to the mid-nineteenth century. The second site is The
Nineteenth Century in Print (hereafter, NCP), a collaborative effort between
Cornell University, the University of Michigan, and the U. S. Library of Congress,
with the goal to preserve American magazines from the 19th century. Both collec-
tions are searchable by keyword using a Web interface, and the results returned
include one or more magazine pages (from the original source) for a single in-
stance of the keyword, thus providing the contextual information needed for close
analysis and for accurate identification of the target emotion.1

1. These features of the study design follow the main study of historical metaphor discussed
in Chapter 4 of this volume.
Chapter 7. The edge of anger

The specific texts chosen for the data collection were two English language
magazines published from 1844 to 1863 Blackwoods Edinburgh Magazine from
ILEJ and Littels Living Age from NCP (hereafter, BEM and LLA, respectively). The
North American Review (hereafter, NAR) was later added to balance the number
of samples for British and American English (see Table 4, below). The magazines
and the specific time frame were selected because the digitized volumes for both
publications are complete for the continuous 20-year period.
The selected magazines were searched by typing the word spleen in the search
box provided by the two Web sites. The keyword typed in the search box took the
form spleen* (without the quotation marks). The asterisk indicates that the key-
word is a wildcard, and instances of the keyword with inflectional or deriva-
tional suffixes (e.g., spleens, spleenful, spleeny) will be included in the search results.
The full text of the digitized magazine volumes were searched. The first 100 hits
in the search results for each magazine between 1844 and 1863 were accepted as
samples for analysis. However, in the two American English magazines, the first
200 hits were accepted. This difference in the minimum hits accepted was insti-
tuted in order to equalize the number of metaphor cases collected from each cor-
pus, for the purposes of comparing the samples from the two English dialects in
the data analysis.
The search results pages were printed out for reference and documentation
purposes because the scanned images of each magazine page could not be searched
using the keyword procedure. The search system only identified the page(s) in
which the keyword occurred, and then each page had to be read in order to find
the specific instance of the keywords use. For this reason, the scanned image of the
original article page in which the keyword appeared was also printed; if the scope
of the keyword use covered multiple pages of text, all pages that covered the key-
word usage were printed. In some cases, the page could not be printed in its en-
tirety due to printed page size limitations; in these instances, the material missing
from the printed page was copied by hand on the printout. Finally, some keyword
hits initially appeared to mark a new case but the scanned pages were the same
ones that were collected in previous hits. In sum, a new case could not be identi-
fied until after the scanned pages for that case were read and analyzed in their
entirety. For this reason, a random selection procedure was not employed to select
cases. This important characteristic of the keyword search system for the NCP
database was a significant methodological issue and was accounted for in the data
analysis procedure.
171 cases resulted from the keyword search procedure. 42 duplicate instances,
which were likely artifacts of the corpora search algorithms, were eliminated; the
remaining 129 cases comprised the study sample. 20 non-metaphorical uses of the
Metaphor across Time and Conceptual Space

keyword were also excluded, including medical references, leaving 109 spleen
metaphor cases.

Data selection

The metaphoric expressions collected employed the word spleen as the source do-
main, with a target domain (either present lexically or implied contextually) that
signifies an expression of emotion. In addition, the context in which each instance
of the keyword appeared was carefully read and evaluated to determine the target
domain; only samples categorized within the target domain of human emotion
were accepted for analysis. Sample 1 shows the problems inherent when analyzing
some of the metaphor samples.
(1) Fair Saint George
Inspire us with the spleen of fiery dragons!
Upon them! Victory sits on our helms. (BEM, January, 1858, p. 131)
The metaphor targets courage, which is a personality trait, not an emotion. In
addition, the amount of context available is inadequate to categorize the target
domain of the metaphor clearly. 15 cases were eliminated due to the difficulty of
determining the target domain.
After these cases were deleted, the total cases remaining for BEM were 51 and
43 for LLA. A total of 94 samples were included in the study dataset. Table 4 below
shows the results of the data selection procedure.

Data analysis

In the data analysis phase, the 94 spleen metaphor cases and their original context
were compared to six properties of embodiment from Lakoff and Kvecses (1987)
study of anger. These properties included (1) the container image schema; (2)
heat; (3) the heat scale; (4) pressure; (5) fluid; and, (6) visible physiological ef-
fects (i.e., skin redness, bodily agitation, interference with visual perception). Each
sample was analyzed for all of the properties. The comparison procedure served as
a rubric to aid the systematic identification of specific aspects of embodiment

Table 4. Spleen metaphor study: Keyword instances, excluded cases, and study cases

Magazine Instances* Excluded Study Cases

BEM 68 18 51
LLA/NAR 61 17 43
Totals 129 35 94
Chapter 7. The edge of anger

in the data, and also to determine if the spleen metaphor is a variation or extension
of the blood metaphor or a separate conceptualization, as was done in the main
study of historical metaphors of anger (see Chapter 6). As stated in the Introduc-
tion to this chapter, the purpose is of this procedure is to determine the content of
the spleen metaphor and whether the analysis in this study supports the results of
the main study.


The following section presents examples of spleen metaphors and their common
characteristics. Sample 2 displays the container, pressure, and fluid properties
described by Lakoff and Kvecses (1987).
(2) In short, altogether he is put out, and he vents his spleen on the swans,
which follow him along the wave as he walks along the margin, intimating
either their affection for himself, or their anticipation of the bread crumbs
associated with his image... (BEM, January, 1859, p. 1)
The context of the sample indicates conceptualized anger in the words put out.
The spleen metaphor is in its prototypical form, he vents his spleen. The container
property in the data focuses on the spleen, rather than the human body.
The properties of pressure and fluid are instantiated in the word vents. The
Oxford English Dictionary Online (OED-O) states that the verb can mean to dis-
charge, eject, cast or pour out (liquid, smoke, etc.); to carry off or away; to drain in
this way... Said usually of the containing thing, but sometimes of the force or
means by which outlet is given. Force indicates pressure on the fluid. An alterna-
tive meaning of the verb applies to ...persons, animals, or their organs: To cast out,
expel, or discharge, esp. by natural evacuation; to evacuate (all above quotations
from OED-O). Therefore, in Sample 2 the fluid is vented from the spleen to reduce
pressure, and this conceptualization denotes a verbal expression of anger.
The following sample also shows pressure in the container.
(3) There is one fallacy, however, still current against woman, which we must
take this public opportunity of renouncing. A certain ungallant old
father, soured by the circumstances of his lot, relieved some of his spleen
by defining women [Greek translation] an animal that delights in
finery...(LLA, May 1847, p. 337)
The use of the verb form relieved indicates that the spleen is under pressure, and
also that the pressure can be decreased by direct action; OED-O describes the
meaning of the word as [t]o give (a person, part of the body, etc.) ease or relief
Metaphor across Time and Conceptual Space

from physical pain or discomfort. The social situation indicates that the man is
taking out his anger concerning his own disappointments in life and directing
his emotion against other people, in particular women in general. The two samples
presented show some productivity in the linguistic forms, with different samples
employing different words to instantiate the same conceptualization.
Sample 4 below shows the negative consequence of failing to relieve the pres-
sure on the spleen.
(4) And this interesting piece of geographical, and geological, and hydro-
graphical meditation makes part in a burst of indignant spleen which is
to go near to annihilating Man from the face of the Globe! (BEM, Au-
gust 1854, p. 201).
Burst denotes a sudden, explosive destruction of the container as a result of ex-
treme pressure; the violent force of the explosion is displayed in the metaphorical
destruction of the Earth. This property of destructive force is similar to the one
displayed in the anger metaphors analyzed by Lakoff and Kvecses (1987).
The idea of bursting was also expressed in the data through various synonyms,
such as ebullition, in Sample 5.
(5) Swift calls Ruvigny a deceitful, hypocritical, factious knave, -a damnable
hypocrite of no religion; but this is a mere ebullition of spleen, such as
was common with Swift against a Whig opponent. (LLA, November 4,
1854, p. 495)
Ebullition in the OED-O is defined as [t]he process of boiling, or keeping a liquid
at the boiling point by the application of heat. The presence of a word which indi-
cates heat is significant, since that is congruent with Lakoff and Kvecses findings
for prototypical anger. A context analysis of the 94 cases collected found that in
four cases words indicating heated liquid or heated containers were used ebullition
was used in three cases, or 3.2% of the total cases; vehement was used once (1.0%).
Therefore, as found in the main study, the heat conceptualization was rarely em-
ployed in spleen metaphors, indicating it is a minor elaboration of the CM.
The collected samples target other emotions besides anger. Sample 6 is an
(6) It would be in vain to describe the manner in which Mr. Darnell vented
this or similar remarks of mocking irony, or sarcastic spleen. It was not
bitter or sneering, but in his usual mellifluous level tone and passionless
tranquility. (BEM, August 1857, p. 138)
The case maps irony and sarcasm to the source domain; other cases targeted
humor and laughter.
Chapter 7. The edge of anger

All of the spleen samples include the fluid concept, but the qualities of the
fluid are often different from the fluid found in Lakoff and Kvecses data though
always under pressure, spleen fluid is generally unheated, with a few exceptions, as
noted previously. Sample 7 provides further evidence of the absence of heat.
(7) ...he passed the next ten years of his life agreeably enough, if not content-
edly. He found a vent for his spleen in the practice of political journal-
ism, and it was during this period that many of his finest works were
written... (LLA, September 12,1863, p. 518)
heat entails rising fluid and visible physiological effects, such as skin redness and
bodily shaking; those characteristics are not found in Sample 7 or in any of the
spleen samples. The four samples of heat also do not display the physiological ef-
fects of heat. Unheated fluid and the lack of visible physical effects are central
characteristics of the spleen metaphor data, in this study as well as in the main
study of historical metaphors of anger.
As with the main study of historical metaphor, the temperature scale was
present in the spleen metaphor data; that is, an increase in emotional intensity is
not the result or the cause of an increase in temperature. Instead, the cold (un-
heated) end of the scale is instantiated. Sample 8 is an example.
(8) Those who knew him best say that, about this time, his temper became
horribly soured. He never had been very agreeable in the servants hall, but
now he was snappish and morose...But as he durst not quarrel with Gray,
he resolved to vent his spleen upon somebody else, and to his own infinite
misfortune, selected Protocol as the victim. (BEM, February 1853, p. 169)
The emotional intensity increases as the passage progresses the man is first dis-
agreeable, then sour, then morose, and finally decides to vent his increasing anger
by quarreling with another man, named Protocol. The fluid pressure increases to
the point that venting is desired, yet the fluid temperature does not change; the
increase in temperature that signals increased anger in the Lakoff and Kvecses
samples is not present. This sample exhibits the same characteristics of the tem-
perature scale that were found in Sample 5. Fluid in the spleen metaphors is
unmarked for heat, except for the four cases discussed previously, and in all cases,
greater intensity of emotion is not the result of increases in heat. The bursting of
the container is caused by excessive fluid volume in the spleen, exerting pressure
on the container, eventually bursting the container.2 These findings support the
findings of the main study for the spleen metaphor.

2. Kovecses (2005) has found similar results relating to pressure on unheated fluid, as dis-
cussed in Chapter 6.
Metaphor across Time and Conceptual Space

In the main study of historical metaphor (see Chapter 6), the heat scale was
renamed the temperature scale because emotions targeted by the spleen meta-
phors were found on the cold side of a scale that extends from hot to cold (or, high
heat to unheated). Metaphors of the keyword blood (and their attendant conceptual-
izations of emotion) are conceptualized on the hot side of the scale, and metaphors
of the keyword spleen are conceptualized on the cold side. The samples analyzed for
this support the finding in the main study concerning the temperature scale.
In line with the absence of heat and the highly variable nature of the tem-
perature scale (compared to the more limited heat scale), the sixth property of
Lakoff and Kvecses (1987) model, visible physiological effects, is also absent in
the spleen metaphors. In fact, none of the collected samples display the visible ef-
fects identified by Lakoff and Kvecses. Two cases illustrate the lack of physiologi-
cal effects. First, in Sample 9, black bile affects the mind, rather than the body.
(9) That those who would die for and with each other in the hour of peril, are
but too apt to misuse the hour of prosperity in conceiving groundless jeal-
ousies, in attributing undue importance to passing bursts of spleen and
petulance, in mutual and self-torment. It is the original sin of man to take
advantage of the absence of important evils to magnify in his imagination
those of minor consequence... (LLA, October 19, 1844, p. 670)
The result of the metaphoric bursts of spleen is mental suffering for both the person
who expresses spleen anger and those around him or her. Mental health is impli-
cated in Sample 9 through the use of emotion words (e.g., petulance) and words
linked to internal thoughts (e.g., mutual and self-torment and magnify in his
imagination).3 However, visible physiological changes are not displayed. Thus, the
bursting of the spleen (i.e., a verbal, non-violent expression of emotion)4 occurs
suddenly and without warning, due to the lack of visible signals of anger. These fea-
tures also support the findings of the main study concerning the spleen metaphor.
Sample 10 shows the extreme effects of spleen sadness on the sufferers mental
(10) When labouring under a bad attack of the spleen so said our volatile and
veracious neighbours the Englishman felt his life to be a burden to him.
Nothing but family considerations...prevented him from blowing out his
brains with a pistol, or effectually ridding himself of his woes by plunging
into the muddy torrent of the Thames. (BEM, September 1861, p. 302)

3. As with samples found in the main study, magnify in his imagination also indicates that
imaginative thinking is caused by the spleen; this follows the unified model of human health.
See Chapter 6 for discussion.
4. Verbal, non-violent expressions of emotion were also found for the spleen metaphors ana-
lyzed in the main study. See Chapter 6 for discussion.
Chapter 7. The edge of anger

The text indicates that prolonged or intense exposure to black bile results in ex-
treme negative thoughts, including the consideration of suicide.5 Visible physio-
logical sensations, such as skin redness, are not manifested in spleen metaphors.


The research questions for the study focused on the structure and meaning of
spleen metaphors, and also on the physical experience that produces the cognitive
conceptualization. The collected samples show that spleen metaphors are system-
atic and consistent in their instantiation of several features found in the main study
of historical metaphor: the container is the spleen, and the container is under
pressure; there is fluid in the spleen, and the fluid is usually unheated; the
temperature scale is present; a variety of emotion types are expressed, including
anger, sadness, madness, and sarcasm, and (mocking) mirth; the expression of
emotion typically occurs suddenly and without warning (due to the lack of visible
physiological effects), and the resulting violent behavior can have severe emotional
and psychological consequences, such as depression and suicidal thoughts, for the
person expressing the emotion and for others present at the time of the outburst.
To summarize the findings for the metaphor analysis specifically for the six
characteristics of Lakoff & Kvecses CM of anger (1987), the spleen metaphors
in this study were found to employ pressure and fluid consistently, but heat was
rarely used, present in four cases out of 94 (4.2%). Two of the characteristics were
changed significantly in the spleen metaphors: the container references the
spleen rather than the human body, and the heat scale extends to cold tempera-
tures. As was done in the main study of historical metaphor, the heat scale has
been renamed the temperature scale. Finally, none of the metaphor cases
analyzed in this study displayed physiological effects of anger. Of the six charac-
teristics, two were employed consistently across the samples, and four were either
significantly changed in conceptual content or were not employed.

Implications of the study

There are several implications of these results. First, the experiential basis of spleen
metaphors is not located in the physical experience of the spleen. The linguistic ex-
pressions consistently instantiate as physical experiences that humans do not ac-
tually experience in the world, including unheated bodily fluid, the ejection of that

5. This feature follows the unified model of human health. See Chapter 5 for further details.
Metaphor across Time and Conceptual Space

fluid from the body, and the perception of the spleen as a source of the ejected fluid.
Spleen metaphors do have a basis in physical bodily experience the experiential
scene of a person expressing emotion but in the historical period under study the
interpretation of that scene originated in the unified model of human health, as dis-
cussed in Chapters 1 and 5 of this volume. As was found in the main study of his-
torical metaphor, this conclusion indicates the importance of cultural models for
interpreting an experiential scene within a particular speech community.
Second, the differences between the diachronic spleen metaphor and syn-
chronic studies of the blood metaphor, discussed in detail in Chapters 2 and 4,
were also found in this study and indicate a significant difference between the two
metaphors. This finding again questions the proper categorization of spleen meta-
phor: do these linguistic expressions represent peripheral, non-prototypical mem-
bers of the cognitive category of anger or a different category altogether?
Answering this question can reveal important aspects of the anger category and
also of conceptualization, specifically the boundaries between categories.
For example, investigating the concept of insanity can delineate important
aspects of the anger category. Lakoff and Kvecses (1987, p. 204) provide many
examples of the conceptual metaphor anger is insanity. An aspect that the au-
thors do not study is the elaboration of the metaphor: insanity can be character-
ized as an illness (italics are from the original):
When my mother finds out, shell have a fit.
When the ump threw him out of the game, Billy started foaming at the
The illness conceptualization extends to madness, which Lakoff and Kvecses
identify as a historical artifact in English:
Im mad!
These observations suggest the possibility of a different category, madness is an
illness, with anger as an elaboration of the metaphor. If this analysis is correct,
then the boundary of the anger category would overlap with illness, indicating
the existence of a complex domain, or domain matrix, of emotion, which incor-
porates all of these conceptualizations as dimensions within the matrix. This con-
clusion supports the findings of the main study of historical metaphor.


This study has found that investigating the boundary of the anger conceptualiza-
tion can provide important details about the concept, as Geeraerts suggested.
Chapter 7. The edge of anger

Spleen metaphors were found to have several properties that differ markedly from
the characteristics of the model of prototypical anger proposed by Lakoff and
Kvecses (1987); the analysis also suggests that spleen metaphors may be better
placed in a different cognitive category and/or in a complex domain matrix with
other conceptualized emotions. The conceptualization of disease and its relation-
ship to anger is an interesting avenue of study in conceptual metaphor because
cultural models of illness have historical significance in English and other European
languages. The implications of this result for the study of conceptualization and
cultural models are discussed in Chapter 9.
chapter 8

Bubbling happiness
Properties of emotion


This chapter discusses the relationship between conceptualization and cultural

models for a different CM of emotion. The purpose is to investigate whether the
complex domains found for anger (Chapter 6) and the spleen metaphor (Chapter 7)
also apply to happiness.
This study is designed to investigate the question whether other emotions that
were instantiated in the main study of historical metaphor would also be found in
a study of emotion metaphor not related to anger. If this result does occur, that
would serve as further evidence of a domain matrix (see Chapter 1) that organized
emotion concepts in historical English.
As a cognitive concept, happiness has several features that appear to instanti-
ate a variety of emotion categories. A sample sentence (from Stefanowitsch, 2004),
constructed by the researcher as a sample of the happiness is a liquid metaphor,
is shown in Sample 1.
(1) She bubbled with joy.
An interesting aspect of the sample is the concept of creating bubbles in a liquid.
Important questions in this care include the following. What does bubbling
mean? Why is it associated with conceptualized happiness? How is it conceptual-
ized via an experiential scene? Does bubbling always express the emotion of hap-
piness? This study will investigate these issues. The purpose is to understand in
more detail the relationship between a specific conceptualization of emotion and
other emotion concepts, a pattern that was seen in the main study of diachronic
metaphor. By this procedure, more details about complex arrays of emotion con-
cepts, as found in a domain matrix will be delineated.


The sources for the data were historical texts in English published between 1500
and 1850. The reason is that the time course of bubbling liquid metaphors may
Metaphor across Time and Conceptual Space

be useful in determining the origin of the form and what conceptual metaphor, if
any, that it instantiates. Historical data is useful for determining the origin and
development of a linguistic form over time. Therefore, the current study will em-
ploy historical data to aid in investigating the pertinent issues. In particular,
electronic historical corpora which can be searched using Boolean or other so-
phisticated search methods will be employed, in order to track more accurately
metaphor over time.
The method for searching will employ what we will call the metaphoric prop-
erty search technique. This corpus search method employs lexical items that in-
stantiate specific properties of the conceptualization, rather than items of the
source or target domain of the metaphor. In emotion metaphors, the source is of-
ten a concrete object, such as blood, but the target is the emotion His blood
boiled. The emotion in this sample is not present lexically; it is an implied target,
as discussed in Chapter 4. It is possible that the standard keyword search method
will not find all instances of happiness, and certainly not all those using bubbling
liquid as a property. Therefore, in order to collect samples that specifically em-
ploy the property under investigation, the search procedure should employ lexical
items that instantiate the property, rather than the target or source. For the current
study, metaphoric property search entails keyword searching using lexical items
attached semantically to the property of bubbling. The keywords employed for
the study were syntactic and semantic variations of the lexical item bubble.
Summarizing this volume, studies of the effects of cultural knowledge on CM
of emotion have found inconsistencies in the metaphoric expressions that do not
reflect cognitive processes; rather, the inconsistencies reflect the influence of cul-
ture. The goal of the current study is to investigate the bubbling liquid property
of the happiness CM to look at the relationship between the property and the
underlying CM. If there is a cultural model that provides perspective on the expe-
riential scene, then this is further evidence for the existence of a domain matrix of
emotion, as discussed in Chapters 6 and 7.

Research questions

Based on the above discussion, the following research questions were investigated.
1. What is the meaning of bubbling in happiness metaphors?
2. Do the properties of bubbling liquid indicate the existence of cultural mod-
els and/or a domain matrix of emotion?
3. Is the metaphoric property search technique useful for collecting metaphor
samples from electronic corpora?
Chapter 8. Bubbling happiness



The metaphoric expressions of bubbling liquid were collected from two online
sources. They were chosen for their coverage of the historical period (1500 to
1850), ease of use, and low cost (both were free for public use).
The first source was the public-access corpus at the Electronic Text Center,
University of Virginia. The Center states that its Modern English Collection con-
tains fiction, non-fiction, poetry, drama, letters, newspapers, manuscripts and il-
lustrations from 1500 to the present, arranged for browsing by authors last name
or by category of interest (University of Virginia, Search page for publicly-accessi-
ble collections, paragraph 1). Sources from British sources and American sources
are included in the corpus. The UVa Etext system was searched from 1500 to
1850 for metaphoric expressions.
The second source was the North American Review magazine in The Nine-
teenth Century in Print digital archive at Cornell University. The archive includes
23 separate magazines published in the U.S. The North American Review collection
covers the years 1815 to 1900, and it is the largest collection of volumes in the ar-
chive. The entire collection was searched for bubbling liquid metaphors.

Data collection

The metaphoric expressions were collected by entering the keyword bubbl in the
search box of the digital archive electronic search systems. The keyword did not
include the word-final e in order to include suffix variants in the results, such as
bubble, bubbled, bubbles, and bubbling. In the Virginia corpus, the search was lim-
ited to 50-year increments, in order to reduce the size of the results list. The corpus
was searched in this manner from 1501 to 1850. The UVa search system listed the
results by work (book or magazine), rather than in chronological order. Also, the
entries in the results list had to be checked by hand because the results list showed
only the citation for the work; the text which contained the keyword was not dis-
played until the results citation hyperlink was selected. As a result, the entries had
to be opened individually and read in order to determine if the use of the keyword
was a linguistic metaphor. In the Cornell corpus, the results were listed chrono-
logically; however, the resulting citations list still had to be checked manually for
metaphoric expressions. This issue markedly slowed data collection in both of the
online corpora.
After finding all of the metaphor cases, the cases were retyped in a word pro-
cessing program and placed in chronological order for analysis. From the UVa
Metaphor across Time and Conceptual Space

corpus, 12 metaphoric expressions of emotion employing bubbling liquid were

found, out of 74 samples which used the keyword; the oldest sample was published
in 1832. The procedure for the North American Review was the same as outlined
above; however, the entire corpus was searched at once. A total of three meta-
phoric expressions were found, out of 60 samples; the earliest was published in
1817. A total of 15 samples of bubbling liquid metaphors of emotion were col-
lected for the study in 134 keyword instances.
Interestingly, the metaphors collected targeted a variety of emotions and per-
sonality traits, including happiness, excitement, inspiration, sadness, and
anger. This result supports the main study of historical metaphor, including the
temperature scale found in the main study of historical metaphors of anger.
These aspects of the results will be taken up in the Discussion section.

Data analysis

The metaphoric expressions were analyzed for their properties using the lexical
items in the text, the lexical items in the surrounding sentences, and the situation-
al context of the passage. In some cases, the emotion target was implied, requiring
the use of context to determine the target domain of emotion. Finally, two tables
were created, comparing the properties of the temperature scale and pressure
to the collected samples. These materials formed the basis of the analysis.


The analysis of the data samples indicates that the property of bubbling liquid is
found in several metaphors of emotion. These metaphors, in order of frequency
in the data, include happiness, excitation, sadness, and anger. The bubbling
property spans across emotion categories, similar to the temperature scale
found in the main study of historical metaphor. Each of these emotion types will
be discussed in turn.

Metaphors of happiness

Of the 15 data samples of bubbling liquid, eight of the metaphors instantiated

happiness. A typical example is shown in Sample 2, from Elinore Stewarts Letters
of a Woman Homesteader (1847).
(2) After we had put the horses in the barn we had dinner and I heard the
story of the girls odd names. The mother is one of those comfy, fat little
women who remain happy and bubbling with fun in spite of hard knocks.
Chapter 8. Bubbling happiness

The italicized words indicate the instantiation of the properties of the metaphor in the
context of the passage: happy and fun indicate happiness, women instantiates the
container metaphor, and bubbling stands for the fluid in the container. In addition,
the liquid in the container is pressurized, in order for the liquid to bubble and flow out
of the container and become visible happiness. Finally, the liquid is unheated (i.e., on
the cold end of the temperature scale). These are the basic elements of the happi-
ness is a liquid in a container conceptual metaphor, as discussed by Kvecses
(2000); these properties are found in six of the eight happiness metaphor samples.
The remaining two happiness samples have the same properties discussed
above, except for the property of pressure on the fluid. A sample of this type is
found in Sample 3, from a poem published in the North American Review maga-
zine in 1817. This is also the oldest sample found in the corpus.
(3) That life is but the summer insects play,
Who breathes, lives, flutters, dies, within a day;
Our happiness, a bubble, and so frail,
It bursts before the Zephyrs slightest gale;
The currents foam, fit emblem of our joys,
That the first ripple scatters and destroys;
That virtue is but seeming hope and love
Glow dimly here, not kindled from above;
Whilst sin and want and misery alone
In this our sphere, have reared a lasting throne.
The italicized words instantiate the properties discussed in Sample 3, but pres-
sure on the fluid is absent. Instead, the foam (a mass of bubbles) on the surface of
the fluid is a result of the movement of waves or currents. The liquid is therefore in
an open container, such as a river, a lake, or an ocean, and pressure is thus absent
due to the absence of an enclosed container. It can be argued that the pressure in
this case is air pressure within the bubbles themselves (causing the bursting), rath-
er than fluid pressure exerted by the liquid from outside the bubbles. This point is
an interesting one, since in Lakoff and Kvecses (1987)s model of anger, pres-
sure is termed internal pressure (Lakoff, 1987, p. 381) without specifically refer-
encing air or fluid pressure. However, the samples given by Lakoff and Johnson are
bodily fluids (Lakoff, 1987, p. 382 and 385); therefore, all of their samples reference
fluid pressure. The other sample for air pressure is shown in Sample 4.
(4) The difference between the adult and the child being only one of degree,
and that but slight, are not bubbles of philosophy as necessary for the one
as bubbles of water for the other? Are not delusions by their very hollow-
ness all the better fitted to buoy up humanity, and float it over the rough
waters until it shall reach the firm soil?
Metaphor across Time and Conceptual Space

Therefore, we posit that the type of pressure instantiated in the happiness meta-
phor is air pressure, not fluid pressure. In sum, the properties found in Sample 3
and 4 diverge from Kvecses (2000). Without the container to create pressure on
the fluid, the property of pressure is absent. Both of the happiness metaphor
samples in which pressure is absent are missing the container metaphor. The
result is consistent with Lakoff and Johnsons (1980) features of the fluid in a
container metaphor without fluid pressure on the liquid, there is little motive
force for the liquid to flow outward or to bubble within the fluid. The property of
fluid pressure is missing in these samples.

Metaphors of excitation

Five samples of excitation metaphors were found in the 15 data samples. This
label is used to refer to the quality of excitement (i.e., churning) of the fluid in the
metaphor, and also to distinguish the category from agitation, a term used by
Lakoff and Johnson to describe the physical shaking of the human body found in
some anger metaphors. The qualities found in the happiness metaphors are also
found in the excitation metaphors. This category of metaphor includes four
inspiration metaphors and one frenzy metaphor. The reasons for putting these
two types in one category is that (1) both types instantiate the same properties in
the conceptual metaphors; and, as mentioned above, (2) the conceptual idea be-
hind both types is excitation of the bubbling liquid. That is, the liquid is
churned or flows rapidly outward as a result of excitation. An example of the inspi-
ration category is shown in Sample 5, from William Godwins Thoughts on Man:
His Nature, Productions, and Discoveries (1831).
(5) He has no resemblance of the art, so conspicuous in Fletcher and Far-
quhar, of presenting to the reader or [s]pectator an hilarity, bubbling and
spreading forth from a perennial spring, which we love as surely as we feel,
which communicates its own tone to the bystander, and makes our very
hearts dance within us with a responsive sportiveness.
The excitement of the fluid is instantiated in the properties of bubbling and flow-
ing, also indicating that the fluid is pressurized. In addition, the metaphorical
inspiration that comes from the flowing liquid of the artists creativity is implied.
Therefore, Sample 5 instantiates the inspiration form of the excitation meta-
phor. As in previous samples, the quality of heat is absent; none of the five
metaphors in the category instantiate heat. The excitation metaphors displayed
the same properties as the happiness metaphors.
Chapter 8. Bubbling happiness

However, one of the excitation samples hints at the possibility of heat. The
case is shown in Sample 6, from Nathaniel Hawthornes House of the Seven Gables
(6) Clifford shivered from head to foot. The wild effervescence of his mood
which had so readily supplied thoughts, fantasies, and a strange aptitude
of words, and impelled him to talk from the mere necessity of giving vent
to this bubbling-up gush of ideas had entirely subsided. A powerful ex-
citement had given him energy.
In this sample, the one word that may instantiate the heat property is effervescence.
In Merriam-Websters Collegiate Dictionary (hereafter, MCWD), the Latin root of
the word means to boil. Is this the meaning expressed in the sample? It is difficult
to determine the answer because the rest of the passage does not provide much evi-
dence of heat. Close analysis of the sample is needed to investigate this question.
Two pieces of evidence argue against the possibility that effervescence means
boiling in this sample. First, as discussed in Chapter 6, vent- is used in both blood
and spleen metaphors, but each type instantiates a different point on the tempera-
ture scale hot for blood metaphors and cold for spleen metaphors. However, as
was found for sample (39) in Chapter 6, the behaviors that result (i.e., gush of ideas)
are not related to the public acts of retribution found in Lakoff and Kvecses
anger Prototype Scenario, a result that was also found in the main study of dia-
chronic metaphor. The sample instantiates verbal (i.e., talk), non-violent acts of
private imagination, similar to the spleen metaphors discussed in Chapter 6. In ad-
dition, shivering instantiates the temperature scale on the cold side of the scale.
Therefore, the evidence does not support the boiling meaning for effervescence.
The happiness metaphors and the excitation metaphors instantiate unheated
fluid, similar to spleen metaphors. This result is consistent with the main study
of historical metaphor (Chapter 6) and the spleen metaphor study (Chapter 7):
the same property, in this case bubbling, can instantiate a range of related emo-
tion states.
Finally, there is one sample in the excitation metaphors which does not in-
stantiate pressure, similar to the two happiness samples discussed above. This
case is shown in Sample 7, from the North American Review magazine in 1832.
(7) Minds that could not understand his beauties, could imitate his great and
glaring defects. Souls that could not fathom his depths, could grasp the
straw and bubbles that floated upon the agitated surface, until at length
every city, town, and village had its little Byron, its self-tormenting scoffer
at morality, its gloomy misanthropist in song.
Metaphor across Time and Conceptual Space

Like the two happiness samples, the quality of excitation (agitated) during in-
spiration is clearly present, placing Sample 7 in the excitation category. However,
as before, pressure is not instantiated, due to the liquid being held in an open
container. Therefore, the absence of pressure is again consistent with Lakoff and
Johnsons (1980) description of the container metaphor, in which a closed con-
tainer creates fluid pressure. The container in this case lacks fluid pressure; the
churning of the surface of the fluid maps to the target of excitation.

Metaphors of sadness

Sample 8, unlike the other samples collected, is a case of an implied target. For
reasons given below, it was classified as sadness. The sample is from Ralph Waldo
Emersons Essays: Second Series (1844).
(8) The street is full of humiliations to the proud. As the fop contrived to dress
his bailiffs in his livery, and make them wait on his guests at table, so the
chagrins which the bad heart gives off as bubbles, at once take form as la-
dies and gentlemen in the street, shopmen or barkeepers in hotels, and
threaten or insult whatever is threatenable and insultable in us.
Sadness is implied in the lexical item chagrins. The liquid in a container meta-
phor is also not clearly instantiated. The bubbles are given off, possibly meaning
that they form on the outer surface of the container (the heart), rather than inside
of it. However, the verb could also refer to bubbles that have escaped the container.
In addition, the fluid is not heated and the container may not be completely en-
closed (or the liquid is on the outside of the container), and pressure is not clearly
Several of the samples, including those for excitation, sadness, and anger
metaphor lack fluid pressure, due to the open or missing container. Yet, these
features are in line with Lakoff and Johnsons characterization of the container
metaphor only a closed container exerts internal pressure on the fluid. The open
container metaphors instantiate properties that would be present in a fluid held in
an open container. This conclusion indicates the presence of a different concep-
tual metaphor. The implications are discussed in the next section.

Metaphors of anger

Like sadness, there is only one sample of anger in the data. The fact that the
metaphor employs bubbling liquid is interesting. The sample is qualitatively dif-
ferent from the happiness, excitement, and sadness metaphors.
Chapter 8. Bubbling happiness

The anger sample is also interesting for its use of an open container, the ab-
sence of pressure, and the employment of the cooking semantic frame, dis-
cussed in Chapter 6 for the main study of historical metaphor. This type is shown
in Sample 9, from Nathaniel Hawthornes book, The House of the Seven Gables
(9) being a pretty warm morning, she bubbled and hissed, as it were, as if
all a-fry with chimney- warmth, and summer-warmth, and the warmth of
her own corpulent velocity. She tried the shop-door; it was fast. She tried
it again, with so angry a jar that the bell tinkled angrily back at her.
anger is instantiated in the heated, bubbling, and hissing fluid. The word a-fry
indicates that the fluid is cooking oil, not blood, and also a fry pan, rather than a
closed container. The word chimney-warmth is also evidence for the open con-
tainer. Excitation of the container is seen in its movement (i.e., velocity) and
violent actions (so angry a jar). The movement of the container in response to
emotion indicates that the sample instantiates the intensity of emotion is
intensity of motion CM found in the main study of historical metaphor. Fi-
nally, consistent with the open container, the metaphor does not instantiate fluid
pressure. Again, other than the violent opening of the door, these properties di-
verge from synchronic studies of the blood metaphor.
In sum, the 15 samples collected for emotion metaphors of bubbling liquid
indicate that there are alterations in the properties of heat, pressure, and the
presence (or absence) of the container image schema in the data. The anger
metaphor displayed heat while the metaphors of happiness, excitation, and
sadness did not; conversely, those three emotion types did instantiate pressure,
while the anger metaphor did not. These results are summarized in Table 4 and
Table 5. The implications of the results are discussed in the next section.

Table 5. Bubbling Liquid Metaphors: heat Properties1

Emotion Metaphor

Property happiness excitation1 sadness anger Totals

No heat 8 5 1 0 14
heat 0 0 0 1 1
Totals 8 5 1 1 15

1. The excitation category includes both inspiration and frenzy metaphoric expressions.
Metaphor across Time and Conceptual Space

Table 6. Bubbling Liquid Metaphors: pressure Properties

Emotion Metaphor

Property happiness excitation* sadness anger Totals

No pressure 2 1 1 1 5
pressure 6 4 0 0 10
Totals 8 5 1 1 15


The three research questions can now be answered based on the above analysis.
The first question investigated the meaning of bubbling in happiness metaphors.
From the data, it appears that bubbling instantiates intensity of emotion is
intensity of motion, a CM found in the main study of diachronic metaphor. The
baby that bubbles with delight and the woman bubbling with fun, in Sample 2,
express visible signs of happiness, including laughter and joking. Other intense
emotions were also targeted, indicating that bubbling liquid, like the metaphors in
the main study, was used to instantiate a variety of emotion concepts.
The second question raised by the study relates to the relationship between the
bubbling liquid property and cultural models. The bubbling metaphors exhibit
characteristics of the unified model of human health investigated in the main study
of historical metaphor. As discussed in Chapter 4, the spleen contained cold black
bile and was associated with laughter and mirth. These same characteristics are
found in the happiness, excitation, and sadness metaphors. In addition, the
temperature scale, found in the main study of historical metaphor, was present
across the four emotion categories found in this study, with anger on the hot end
of the scale and the other three emotions on the cold end. Finally, as with the
spleen metaphors in the previous chapter and also the main study of historical
metaphor, the bubbling metaphors, with their variety of emotion targets and the
cross-domain temperature scale, point to the existence of a domain matrix of
emotion that relates all of these emotion categories in a complex cognitive system
of physical experience in the world and cultural models. This conclusion is dis-
cussed in detail in Chapter 9.
The third research question, which asked whether the metaphoric property
search procedure is useful for collecting metaphor samples, is rather complex. The
technique was able to identify samples of metaphors in which the target emotion
is not lexicalized in the text. The technique showed clearly the cross-domain na-
ture of the bubbling property, which was the major goal of the procedure. Though
Chapter 8. Bubbling happiness

the technique shows some potential for metaphor research, further studies are
needed to investigate its usefulness as a research method.

Conclusions and implications of the study

The implications of the metaphoric property search technique for the current re-
search are clear, however. The procedure cut across the lines of emotion meta-
phors, identifying samples for four separate emotion domains happiness,
excitation, sadness and anger. The result supports Kvecses (2000) survey
analysis of emotion CM. That analysis made it clear that different emotion meta-
phors share properties and employ some of the same conceptualizations. There-
fore, searching an electronic corpus using a keyword for a lexicalized metaphoric
property, rather than a source or target domain, has the unique ability to identify
samples which share that property, cutting across different domains and providing
details about the conceptual relationships between those domains. For research in
metaphor properties or for survey research of complex domains, the metaphoric
property search technique for electronic corpora may prove to be a useful one.
Further research on this technique is recommended.
One more important implication of the study involves the creation of a new
conceptualization. As was noted in the data collection phase, the corpora were
searched between the years 1501 and 1899, but the earliest sample of a bubbling
liquid metaphor of emotion was found for 1817. This particular sample was one
of the two happiness metaphors that did not instantiate pressure.2 The other
bubbling metaphors, including all of the ones instantiating pressure, date from
1831 and later. The result points to the possibility that bubbling liquid meta-
phors of emotion are relatively new in English history. While the evidence is at
best circumstantial, further diachronic study may reveal important details about
the origins of the conceptualization and its relationship to historical events, social
norms, and cultural knowledge of the time period.

2. The other happiness metaphor that lacks pressure is dated 1879.

part iv

Conclusions and implications

chapter 9

The non-autonomous nature of cognition,

language, and culture


The chapter is divided into three sections, beginning with a review of the main
studys research questions. Next, the results of the main macrostudy are dis-
cussed concerning the research questions. The impact of the research line on CM
theory and research methodology is discussed next, including recommendations
for future research.

Research questions

The research questions for the main study are reviewed below.
1. What was the conceptual relationship between the blood and the spleen dur-
ing the historical period? Are they located in the same CM, different CM, or is
the relationship characterized in some other way?
2. What motivated the conceptualization in each type of metaphor? Is it bodily
experience, cultural knowledge, a combination of these, or some other source?
3. Did changes in cultural models in turn change the cognitive conceptualization
of anger over time?
4. Did scientific knowledge (and advancement in that knowledge) influence the
cognitive conceptualization of the CM of anger?
These questions will be answered in turn, referencing the results of the ancillary
study of historical sources and the main study of the CM of anger.

Question 1

Based on the results of the frequency study and the discourse analysis, during the
historical period under study, the blood and spleen metaphors investigated here
Metaphor across Time and Conceptual Space

were separate prototypes.1 Each had its own conceptual structure and semantic
meaning, indicating that each had its own experiential scene, conceptual domains,
and cognitive mappings. In addition, both were employed to express different
types of anger and emotional behavior the blood metaphor instantiated an expe-
riential scene of public, violent anger; the spleen metaphor employed the scene of
non-violent anger which could be expressed either publicly or privately. This result
may further suggest that the blood and spleen prototypes were in complementary
distribution; that is, each concept was selected for use in different discourse situa-
tions to communicate different messages. These results support the findings of
other recent studies in metaphors of anger and blood, including Geeraerts,
Gevaert, and Speelman (2011) and Sim (2011). However, this conclusion for
complementary distribution requires further study to corroborate its validity.
Based on these findings, the two prototypes are within the same conceptual
category; however, it is not the CM of anger or any single domain. We propose
that the blood and spleen prototypes are both part of the domain matrix of emo-
tion (hereafter DME). The DME was the basic conceptualization for all types of
human emotion during the historical period up until at least the late 19th century;
the matrix includes non-autonomous, encyclopedic knowledge of the human body
embodied experience and shared cultural models, including scientific knowl-
edge which are organized in a complex system of conceptual relations.
The discourse analysis of the metaphor samples shows several aspects of the
DME. First, the DME conceptualizes emotion rather than anger because many
different emotions are instantiated in related metaphoric expressions. The emo-
tions included anger, hate, love, fondness, joy, calmness, envy, fear, grief,
happiness, excitation, and sadness. The number of emotions instantiated by
the samples and their close relationship to each other (see below for discussion)
indicates a complex domain that organizes these domains of emotion. There are
also a number of elaborations, extensions, and entailments. These include desires,
such as revenge, grudge, sexual attraction, and ambition, and personality
traits, including imagination, intelligence, and reason. The Domain Matrix
of emotion was a complex combination of a wide variety of domains. The DME
was the basic category, and it organized the less complex domains in a system of
relations that were employed to instantiate metaphoric expressions that involve
human emotion.
Second, the domain matrix had several dimensions; the dimensions included
temperature, reason, and control. The dimensions organize the emotions,
desires, and personality traits in the unified model in relation to each other with

1. As discussed in Chapter 2, this conclusion for anger supports Kvecses conclusion con-
cerning multiple prototypes for happiness (1991; see also 2010b).
Chapter 9. The non-autonomous nature of cognition, language, and culture

multiple conceptual links. Thus, heat is linked to both anger and impulsive-
ness, and cold is linked to sadness and calculating. Other domains also were
found in the DME, including intensity of emotion is intensity of motion,
which relates the experience of an emotion to physical movement, activity, and
loss of control. Overall, the dimensions and the emotions are linked systemati-
cally in order to motivate metaphoric expressions which employ the links to com-
municate the conceptualization and employ cultural models the unified model
to express and interpret the experience. The DME provided the systematic rela-
tions between domains and cross-domain dimensions, to create complex linguis-
tic expressions in a form/meaning pair that fit the shared common ground of the
English-speaking community of the historical period.

Question 2

The DME is a complex cognitive conceptualization which employs both embodied

experience and cultural knowledge as constituted by cultural models. Features of
embodiment, such as the container and pressure, were found in the collected
data; as well, features of the unified model were found, including the association of
the spleen with sadness and a calculating personality type and the association of
blood with anger and impulsiveness. We argue that the cultural models provide
contextual common ground in a particular speech community, similar to the way
that embodied experience provides experiential grounding. Both types of ground-
ing are necessary to communicate effectively.
The DME provides the systematic organization of the various emotions and
dimensions to meet the needs of the communicative situation. Such combinations
may contradict either embodied experience or cultural knowledge, for the pur-
pose of expressing a new meaning useful for the current situation. For example,
sample (27) in Chapter 6 extends the property of burning black bile found in
spleen metaphors to blood to indicate extreme anger. This extension of anger
contradicts the unified model because blood, with its wet quality, could not burn
(black bile could burn in the medical condition called melancholy adust because it
had the quality of dryness). Though the entailment violates the cultural knowledge
of the unified model, it extends the temperature scale to communicate a
(relatively) higher level of anger in the metaphorical expression than would be
possible in a blood or spleen metaphor alone; it is the combination of blood and
burning that extends anger to a higher level of intensity, and the extension would
be understood to readers because the unified model was well known to them.
Thus, we argue that the writer creatively combined aspects of embodied experi-
ence with principles of the unified model to create a meaning that would meet the
communicative goal of expressing an extreme level of anger not possible in a
Metaphor across Time and Conceptual Space

typical blood or spleen metaphor. In sum, the DME contains several dimensions
and a variety of emotions which can be employed to express highly complex and
subtle shades of meaning.

Question 3

As shown in Chapter 6, the conceptualizations of emotion vary in frequency over

time; the occurrences of the keywords increase during the height of the unified
models popularity, and decrease as scientific advances refute important parts of
the model, with three of the keywords decreasing to zero instances by the 1850s.
The relationship between frequency of use and the historical popularity of the uni-
fied model was the main reason for using the compiled corpora for data collection.
The corpora, which are designed to be representative of the historical use of
English, indicate that the frequency of use of the keywords (and likewise the meta-
phorical use of the keywords) changed as the cultural value of the unified model
changed. The frequency of use changes coincide diachronically with changes in
the scientific validity and the cultural value of the unified model. Therefore, it is
possible that the decrease in frequency of use of the metaphoric expressions indi-
cate a change in the cultural value of the unified model; the relationship between
the model and metaphor use provides some evidence that changes in cultural
knowledge affect the frequency of use of language over time.
The conceptualizations also vary in their use of different cultural models across
different time periods. The unified model was used extensively in the Early Modern
period of English, and specific details drop out over time or are replaced with new
combinations of DME dimensions, such as combining blood with burning to cre-
ate an extreme level of anger that was not possible without the combination.
The cooking semantic frame also was used with increasing frequency after
1850, indicating that it was not a minor elaboration but an important semantic
frame for the expression of intense emotion, especially anger. The existence of
more than one perspective to encode emotion semantics is similar to the existence
of alternate perspectives of an experiential scene to encode deictic orientation in
syntax. These alternate ways of viewing emotion indicate the role of cultural
knowledge in choosing a perspective that fits the shared knowledge of the speaker
and hearer, as well as the needs of the communicative situation.

Question 4

Finally, changes in scientific knowledge appear to correlate with changes in the

DME. In the 15001549 period, metaphoric expressions using the keywords were
not found in the Penn-Helsinki and ARCHER corpora, which were designed to be
Chapter 9. The non-autonomous nature of cognition, language, and culture

representative of English use of the time period. The only samples for the use of the
keywords in corpora were found in the University of Virginia corpus, in treatises
by experts like Calvin and Machiavelli; however, the Virginia corpus was not de-
signed as representative of English usage. The absence of the metaphors in the rep-
resentative corpora could possibly be attributed to the fact that the unified model
was new to English speakers in the early 16th century, having been introduced into
the language in the mid-1400s via France (Gevaert, 2002). Therefore, knowledge
of the unified model may not have spread to large numbers of non-expert English
speakers in the early 1500s.
As the scientific theory spread, from researchers to doctors to patients, use of
the DME and metaphoric expressions could have increased gradually, peaking in
frequency in the Penn-Helsinki corpus during the 17th century. Harveys discovery
of blood circulation in 1628, which refuted the unified model view of humors held
in bodily organs until needed, occurred less than 75 years before the decline of the
frequency of the metaphoric expressions in the 18th century. By the mid-nineteenth
century, three of the four keywords were no longer frequent enough among English
speakers to register in the ARCHER corpus. This is the same period that Virchow
published Cell Pathology, refuting the dyscrasia theory which began in the Greek
humoral system almost 2,500 years before. Overall, the frequency of use of the
keywords in metaphoric expressions over the course of the historical period fol-
lows the historical rise and fall of the popularity of the unified model generally, and
this result suggests that the spread of scientific knowledge among non-expert
speakers in a speech community affected metaphoric meaning and use.
As discussed above, the study of historical metaphoric expressions of anger
found that emotion concepts were bound together by a series of dimensions, in-
cluding temperature, reason, and control, and the relations between the
concepts and the dimensions constituted the complex Domain Matrix of emotion.
In addition, several complex CM were found, including intensity of emotion
is intensity of motion, controlled response over time, and intense
response over time. The DME accounts for the results of the main study of met-
aphoric expressions as well as the two micro-studies and showed that emotion
concepts were highly interrelated at a conceptual level during the historical period
until at least1850. In addition, the studies show that culture is an important factor
in conceptualization, providing choices among cultural models for creating un-
derstanding between speakers and hearers who share those models and allows for
new combinations of conceptual dimensions to create new metaphoric meanings.
The studies also suggest that cultural models are an important factor in diachronic
changes in conceptualization over time. Finally, scientific knowledge has a role in
changing cultural perspectives of embodied experience, resulting in changes in
conceptualization over time.
Metaphor across Time and Conceptual Space

Implications for conceptual metaphor theory

Semantic shift and frequency of use

The study has important implications for conceptual metaphor theory (CMT).
First, cultural knowledge has an effect on conceptualization diachronically. Two
pieces of evidence support this hypothesis. For one, the selection of a particular
perspective on an experiential scene by cultural models, as was discussed in Chap-
ter 1 in regard to deictic orientation, implies that the possible selections will change
over time as cultural models change, leading to changes in syntax and semantic
meaning to fit the new model. The choices that are available cannot remain static
since cultural change will change the specific details of the choices that are avail-
able for creating linguistic structure and meaning, and new choices will lead to
new cultural ideas, as well. This conclusion supports Enfields (2002) constructs of
event typicality, cultural representations, and changes over time in those represen-
tations via their currency within the speech community. Crofts (2008) semiotic
triangle of form, meaning, and shared community common ground is a paradigm
which speaks to the same issues.
One argument against the frequency/cultural model hypothesis is that the se-
mantic meaning of the keywords simply shifted over time, and as a result the his-
torical metaphorical meanings were used less frequently. Semantic shift in word
meaning is a well-documented process in linguistic research, so it will not be re-
viewed here; suffice it to say that the argument needs to be addressed. Our current
response is that meaning shift is the result of many factors, including changes in
conceptualization and cultural models, as Trim (2011) also asserted.

Cultural models as conceptualizations

Therefore, the second argument for the effect of cultural knowledge on changes in
conceptualization over time is that, as the current study showed, different concep-
tualizations of emotion were employed in the metaphoric expressions analyzed,
including the unified model and the cooking semantic frame, and the dimen-
sions of temperature, reason, and control were combined with emotion in
ways that fit the communicative needs of the moment. These variations can create
new semantic meanings which change the meanings of lexical items in the speech
community and in turn these meaning shifts affected the frequency of use of those
lexical items. In sum, the shift in semantic meanings of the keywords over time has
many causes, and we argue, on the basis of the studies described here, that cul-
tural models, as one type of non-autonomous knowledge, are an important cause.
Shifts in semantic meaning can be explained as part of the general process of
Chapter 9. The non-autonomous nature of cognition, language, and culture

conceptual change via changes in cultural models that the results of this study sup-
port. However, the current study was not designed to investigate diachronic se-
mantic shift. Future research needs to address this important issue.

Frequency of use measures and conceptual metaphor

Related to the above considerations is a theoretical principle that has influenced

conceptual metaphor research design for many years, discussed in Chapter 1: the
frequency of use of a linguistic form has little impact for describing the cognitive
concept(s) that motivate the form, especially the prototypicality of the conceptual-
ization (e.g., Kvecses, 2008, p. 200), following the theoretical principle that a con-
ceptual metaphor is a construct of the human mind rather than a linguistic form
(Lakoff, 1993). A wide variety of linguistic forms may express a single conceptual-
ization, and the frequency of use of these forms will not aid the determination of
the concepts prototypicality.
Understanding this principle, in the main study of historical metaphor, the
prototypicality of the blood and spleen metaphors was investigated by analyzing
the experiential scenes that motivate the conceptualizations: if the two metaphors
employ different experiential scenes, this would indicate that each metaphor is a
separate conceptualization, denoting different types of anger. Based on this crite-
rion, the two metaphors were determined to be separate, prototypical conceptual-
izations for anger. Conceptual metaphor research, including the results of the
current studies, supports the theoretical view that frequency of use statistics are
not useful for explaining the prototypicality of a conceptual metaphor.
However, as demonstrated in the main study of historical metaphor, frequen-
cy of use statistics can track the employment of lexical items that indicate the
cognitive mappings between the source and target domains, and also help to
identify the cultural models that provide perspective on the experiential scene.
Introspective data for a historical form employed for these specific purposes
would be highly inaccurate, due to the inaccuracies inherent in the contemporary
researchers own non-autonomous knowledge of the historical form. This was the
primary justification for the use of non-linguistic historical data and frequency
statistics in the main study. Diachronic studies can usefully employ frequency of
use statistics to track the lexical items that instantiate the source domain and its
potential target mappings, as well as aid in the identification of the extant cul-
tural model(s). The results of the current studies demonstrate these important

2. We also argue for the use of frequency statistics for synchronic studies of conceptual meta-
phor, for a different methodological reason. See Chapter 1, Footnote 26.
Metaphor across Time and Conceptual Space

Is construal a dynamic process?: Revisiting the issue

A second implication is that construal may be a dynamic and emergent process,

following a central tenet of usage-based theory, a topic initially discussed in
Chapter 1. The idea that universal concepts are stable across speakers and lan-
guages has been a fundamental assumption of CF research. For example, in Lakoff
and Kovecses (1987) study of the CM of anger, the elaborations, extensions, and
entailments are presented as connected systematically in a static, stable domain.
The extended structures could only be identified and described on the assumption
of structural stability.
Langacker (1994) argues somewhat differently that cognition is a dynamic,
developmental process situated in a culture (p. 28); however, in his formulation,
culture and language are factors within cognition, thus giving embodied, senso-
rimotor experience (i.e., purely psychological factors in Langackers terms) the pri-
mary role in conceptualization. Over time, he argues, culture and language grow
in influence on cognition and construal, as knowledge in these areas grows. In ad-
dition, Langacker posits the process of stratification on construal, in which we
interpret the same basic phenomenon in terms of alternate cognitive structures
representing different levels of organization and cultural influence (1994, p. 28).
Thus, in the concept of time, humans initially have a basic level, embodied, pre-
cultural experience of time; later, at a more complex level, the spatial metaphor of
time is construed, influenced by both universal embodiment and cultural notions,
since space has societal value. Finally, at the most complex (i.e., highest) level,
time is construed with culture-specific metaphors, such as the Father Time per-
sonification. Therefore, Langackers theory of dynamism provides an important
and increasingly influential role for culture in conceptualization over time and at
higher levels of construal; however, the basic level of cognition is reserved for em-
bodied experience in the world, and embodiment has a constant, foundational
role in concept formation.
In a 2002 article, Langacker stresses the point that while languages do clearly
differ due to cultural differences, I will caution against overestimating the extent
of cross-linguistic variation...linguistic diversity can be seen as developing from a
shared array of resources reflecting universal aspects of the human body, mind,
and experience (Langacker, 2002, p. 138).3 Moreover, Langacker ascribes a cer-
tain amount of dynamism to the process of conceptualization, in that cultural
knowledge is increasingly influential over time; however, the core of a cognitive

3. The current studies support Langackers characterization of a shared array of resources;

however, it is argued here that one of those shared resources is the cultural model that provides
the perspective on physical experience that is shared by speakers in a particular speech
Chapter 9. The non-autonomous nature of cognition, language, and culture

concept is stable, based in the cognitive foundation of embodied experience in the

world. Though concepts may become increasingly complex due to the increasing
influence of culture, the basic structure of a concept is embodied experience.
Langackers formulation can be characterized as providing an embodied core
for conceptualization, a stable cognitive structure onto which additional concepts
are added as the cultural periphery in the form of elaborations, extensions, and
entailments over time.4 In sum, the theory is dynamic and variable for the influ-
ence of culture and language, but it is built on a static and invariable structure of
embodied experience.

The embodied core and the cultural periphery

Overall, Langackers (1994; 2002) dynamism is similar in its implications to

Kovecses (2005) idea of potentially universal metaphor, discussed in Chapter 2.
Grady (1997) and Lakoff and Johnson (1999) similarly depict the differing contri-
butions of embodiment and culture by assigning the embodied core to primary
metaphors and cultural models to complex metaphors. The major theorists in cog-
nitive-functionalism and conceptual metaphor provide for culture as a factor in
their theories of cognitive conceptualization; however, similar to Chomskys gen-
erative grammar theory (see Chapter 1), there is a clear dividing line between the
embodied core and the cultural periphery.
The immediate question is how to account for metaphors of emotion which
seem to have cultural knowledge within the embodied core, such as the internal
organs found to instantiate emotion in Chinese metaphors of anger (Yu, 1995), the
stomach as the locus of anger in Japanese (Matsuki, 1995), and the spleen in the
spleen metaphors studied in this volume. Yu attributes the use of internal organs
to traditional Chinese medical practice, a source of cultural knowledge; Matsukis
description of the unknown substance that comprises hara (which means belly)
and its ability to rise to the head during the experience of anger shows that hara is
a Japanese cultural concept, rather than an embodied experience. In sum, both Yu
and Matsuki found embodied concepts that originated in cultural knowledge, not
physiological experience. The same result was found for the spleen metaphor of
the current studies. If embodiment is the core of conceptualization, then why is
embodiment sometimes hidden or not instantiated at all in metaphor, in favor of
cultural models? This is an important question which has a major impact on cur-
rent conceptual metaphor theory.

4. Langackers formulation parallels the constructs of core and periphery in Chomskys gen-
erative grammar theory. See Chapter 1 for discussion.
Metaphor across Time and Conceptual Space

The embodied core and the cultural periphery: An example

Langacker (2002) explains the lack of instantiation of embodied concepts in gram-

matical construal in a comparative analysis of locative phrases in English and
Chalcotongo Mixtec, an indigenous language of Mexico. In a description of the
differences between English prepositions of location and Mixtec locatives, which
are communicated contextually and not lexicalized in a word form or affix,
Langacker states that
...linguistic meanings are elaborate conceptualizations any facet of which may
participate in the correspondences defining grammatical constructions. A rela-
tionship inherent in a component can thus be exploited in grammatical and se-
mantic composition even if it is not singled out as the profile of any constitutive
element (p. 150).

In other words, a feature of embodied experience is still part of the conceptualiza-

tion, even when the language form does not explicitly employ the feature.
Langackers principle fits within the theoretical assumption that conceptualiza-
tions are stable, long-term cognitive structures in the mind which do not change
over time; however, if that fundamental assumption is incorrect, then the idea of
the embodied core does not work, either.
Moreover, the absence of an embodied feature does not in itself explain the
ability that peripheral cultural knowledge has, at times, to take precedence over the
embodied core in metaphor, as Yu (1995), Matsuki (1995), and the current studies
found. Under what conditions can culture override (Kvecses, 2005) embodi-
ment in the instantiation of concepts in metaphor? Is there a principle that governs
how and why the override occurs? The basic issue concerns the factors that ac-
count for variation in conceptualizations, both within and across languages.

Variation in conceptualization

Kvecses (2005) discusses these factors in detail. He argues that there are two basic
sources of variation: (1) differences in experience; and, (2) differences in cognitive
preferences and styles (p. 231). The second type is the result of the degree to which
elaboration and other cognitive processes are applied in different languages (p. 246);
that is, the variable application of cultural knowledge. He then presents lists of fac-
tors that influence variation for the two major sources of variation. The variable
factors for (1) include the physical environment, social context, communicative situ-
ation, history (both group and individual), and concerns/interests; each of these has
sub-categories, as well, such as power relations for social context and physical set-
ting for communicative situation. The factors for (2) include experiential focus
Chapter 9. The non-autonomous nature of cognition, language, and culture

(i.e., aspects of embodiment that the person cannot ignore for cultural reasons),
viewpoint preference (similar to deictic orientation in Chapter 1), prototypes and
framing, and metaphor versus metonymy preference (i.e., different cultures generally
prefer one over the other in linguistic expressions). He discusses each of the types of
variation and concludes that the lists are not in any way exhaustive, and that more
research is needed to identify and investigate the causes of variation more fully.
Kvecses (2005) fully acknowledges the wide variety of ways in which culture
influences and even changes conceptualization, and provides an example, called
the society is a family CM. Americans in the conservative and liberal political
groups differ on the source-target mapping of the CM, resulting in differing views
concerning whether President Clinton should have been impeached for marital
infidelity, and also differing views in the U.S. and France on the issue of marital
fidelity for heads of state. In the U.S., this was an important political issue, and in
France, it was not, due to different mappings within the society is a family CM.
Kvecses concludes the discussion of the example by stating that
More generally, we can perhaps conclude (despite the admittedly contrived nature
of the example used for illustration) that a well-embodied metaphor may make
use of differential mappings (even within the same culture) because of the influ-
ence of the broader cultural context. Thus, we get cases in which social-cultural
experience overrides embodiment (p. 292).

On the same page, Kvecses explains his use of the word override. He asserts
The talk of overrides may seem to suggest some kind of temporal and causal pro-
gression from a universal base to a cultural overlay and an ontologically most
basic part from which other things emerge, or develop. I do not intend any such
interpretation...I view the emergence of metaphor to be simultaneously shaped by
both embodiment and culture (and most likely also by communicative context). I
simply use override as a convenient way of talking about certain incoherencies
and conflicts among the heuristically postulated systems (p. 292).

There are two important points to be made about Kvecses argument. First, he
agrees that culture has an important a role equal to embodiment in the instantia-
tion of conceptual metaphor. Second, he argues that some conceptual metaphors
are potentially universal (p. 293). Though Kvecses accepts the notion that con-
ceptual metaphor is simultaneously subject to embodied experience and cultural
knowledge, he sees some individual conceptualizations as possibly not subject to
cultural knowledge due to their presence across various languages and cultures.
The question posed in Chapter 2 still applies: under what conditions will a meta-
phor be deemed universal? Kvecses offers the same two criteria that other CM
Metaphor across Time and Conceptual Space

theorists postulate (1) embodied experience and (2) the presence of the same
CMs across languages.

Variation in historical metaphors of emotion: The DME

In contrast, in the current studies, the process of conceptualization was dynamic

and variable in some aspects, including embodied experience, and this result
points to the perspective provided by the unified model. For example, in the Do-
main Matrix of emotion (DME), many emotion concepts were instantiated by the
blood and spleen metaphors, spread across both sides of the temperature,
control, and reason scales in the metaphor samples, and these were often
related to concepts found in the unified model. In addition, the closed container
image schema slowly disappeared over time; by the late 1800s, it had become an
open pot in the cooking frame, until in one 20th century sample, the pot disap-
peared altogether. The scales also showed some dynamism; for example, unlike
Lakoff and Kovecses (1987) analysis, both hot anger and cold anger were
found in the samples, and in similar numbers. Both types had positive and nega-
tive connotations. Hot anger involved increased passion and lack of control, and
resulted in violent acts against those close to the person experiencing the anger,
but there were also acts of warmth and kindness. cold anger was associated with
reason and intelligence and the ability to control oneself, but also resulted in vio-
lent acts against others in extreme cases. Thus, during the historical period under
study, especially before 1900 A.D., control depended on the unified model and
the variable nature of context for its instantiation. As well, physiologically, spleen
anger led to a type of extreme hot anger when cold and dry black bile was burned
by heat; yet blood, despite the qualities of heat and wetness, could also burn in
times of extreme anger. The features of embodiment varied across contexts and
time; they were not static and stable, but dynamic and variable, often changed by
the perspective provided by the unified model employed as a cultural model.

The unity of cognitive domains

The description of the DME in the main study of historical metaphors of emotion
parallels a study by Croft (2003/1993) on cognitive domains. The article posits the
idea of the unity of a domain matrix in grammatical constructions, metaphor, and
metonymy. Croft concludes the article with a discussion of the types of unity which
domain matrices display. The three types of unity are (1) domain; (2) mental
space (including physical space and time, p. 200); and (3) selection (grammatical
categories including person and number). Croft adds the caveat that situational
Chapter 9. The non-autonomous nature of cognition, language, and culture

context also influences the construal of meaning (p. 202203). The point of the
discussion is that
[i]n comprehending an utterance, the listener assumes the unities of the domain,
mental space, and selection, and attempts to interpret the sentence as conforming
to those unities...The listener will generally try as much as possible to adjust the
meanings of the parts to yield a coherent interpretation of the whole (p. 202).

An implication of the above statement is that all of the factors discussed in the
Croft study are necessary for construal: conceptual structure, physical experience
in space and time, grammatical structure, and situational context (including cul-
tural knowledge of the situation). In addition, the listener/reader dynamically
adjusts the parts to make sense of the unified whole. In the same way, we argue
that the DME was applied as a conceptual unity (Enfield, 2002) to metaphoric
expressions of emotion, and the parts of the domain (as currently known and
understood by the listener/reader) were adjusted to make the expression coher-
ent in the situational context. Conceptual structure included all of the current
encyclopedic knowledge of the speaker and the hearer, including cultural models,
and the domain matrix structure was applied to construe the meaning of utter-
ances in context.
The current studies support Langackers characterization of a stable core of
physical experience and the shared array of resources; however, it is argued here
that one of those shared resources is the cultural model that provides the perspec-
tive on the experiential scene. Moreover, in the model proposed here, the influence
of cultural models extends to the core of the conceptualization, via Crofts (2003/1993)
unity construct, to provide the perspective on a experiential scene that can be un-
derstood by members of a particular speech community. In addition, Kvecses
(2005) view that a CM employs both embodied experience and cultural knowledge
is supported by the current studies. However, the universal aspects of the concep-
tualization include both embodied experience and the application of cultural
models to interpret that experience. If cultural models are required to interpret
coherently a physical experience in a speech community, as Croft (2008) has ar-
gued, then conceptualizations that are culture free do not meet the requirements
of communication within the community; all conceptualizations (or at least, com-
plex arrays of concepts like the DME) include cultural models.5
If this conclusion is correct, then including two separate and complementary
types of knowledge embodied core and cultural periphery does not accurately

5. This view also suggests why language learners have difficulty learning metaphors in a sec-
ond language; the extant cultural model obscures the universal aspects in ways that require
learning the cultural model in order to comprehend the perspective on the universal experien-
tial scene. See the Epilogue for more discussion.
Metaphor across Time and Conceptual Space

describe the process of conceptualization. In general, the principle of non-

autonomous knowledge would suggest that such a separation does not character-
ize the way in which cognitive concepts are formed. In addition, the current stud-
ies indicate that the historical Domain Matrix of emotion included both
embodiment and cultural models in a large, multi-dimensional array of cognitive
concepts. The implications of the DME and the construct of non-autonomous
knowledge indicate that separating knowledge into core and periphery does not
characterize accurately cognitive conceptualization or its instantiation in linguis-
tic metaphor.

Conclusion: Conceptualization is static and dynamic

Considering the discussion above, conceptualization can be viewed as both static

and dynamic. The features of embodiment vary in different linguistic expressions
because they employ both embodied experience and cultural knowledge simulta-
neously; that is, humans have embodied experience in the world, but that experi-
ence requires interpretation by cultural values and current context at the level of
the CMs embodied core, in order to provide a semantic meaning that can be com-
municated coherently in a specific discourse situation. Embodiment provides the
raw, neutral input, and culture provides the intersubjective interpretation for the
current situational context and shared common ground, in the same way that cul-
tural models choose a specific deictic orientation from among neutral possibilities
in a conceptualized scene. The inseparable nature of embodied experience and
cultural models is a key finding of the current research work.

Speaker/Hearer interaction and the override

As well, the many factors identified by Kovecses (2005) which contribute to varia-
tion, including situational context, have an effect on the features that are chosen
for a particular use of a metaphoric expression. Therefore, we argue that features
that are not present in a particular linguistic expression are neither available nor
hidden they are irrelevant to the semantic meaning that the speaker wishes to
express. Again, the implication is that conceptualizations are stable and dynamic,
employing both meaning-neutral embodied experience and culturally-licensed
knowledge to create new meanings online to fit the current discourse situation.
However, we do not suggest that embodiment and culture do not constrain con-
strual; they do, but the speaker can override them (i.e., make adjustments, as Croft
2003/1993 argues) at any time, depending on the needs of the situation. Thus, the
ability to override belongs to the speaker and the hearer, not to an embodied core
or to cultural knowledge. In sum, embodiment and culture are employed as tools
Chapter 9. The non-autonomous nature of cognition, language, and culture

to communicate the meaning that the speaker chooses to convey, in order to meet
the demands of common ground in a communicative context.

Implications for the current study

Applying these conclusions to the current study, the historical writers of meta-
phoric expressions of emotion employed the common ground of scientific and
cultural knowledge of the unified model of health, knowledge of embodied experi-
ence, and the communicative context to create meanings that would be under-
stood by the typical reader as a member of the speech community at the time of the
communication. These meanings were dynamic in that cognitive categories high-
lighted aspects of the body not actually experienced bodily (such as the spleen), or
new categories were created to provide a coordination device that the reader would
recognize in the context. For example, the use of the cooking frame in the
boil- samples of the late 19th century provided the heat concept that was needed
to instantiate anger; cooking was used (though not actually experienced bodily)
instead of the unified model because by that time the model was no longer a nor-
mal belief (Green, 1995) and therefore not a part of the common ground of the
writer and reader.
Though such meanings are constructed offline in writing, compared to the
online procedure employed in speaking, the general cognitive processes underly-
ing the creation of meaning in both speaking and writing are presumed to be sim-
ilar. Characterizing the creation of meaning in communication as both a static and
dynamic cognitive process provides both the stability and the flexibility needed in
both modes of communication to meet the demands of communication in the
currently stable, shared common ground of the speech community, in a situation-
al context that changes moment-to-moment at least, in the mind of the speaker/
writer and the listener/reader.

Future research in cognitive-functionalism and conceptual metaphor

The conclusions discussed in the previous section have several important implica-
tions for new research in cognitive-functionalism and conceptual metaphor. These
implications are discussed below.

Dynamic construal

First, as discussed above, research in conceptualization in cognitive linguistics

should investigate more fully the concept of dynamic construal. Croft and Cruse
Metaphor across Time and Conceptual Space

(2004) have developed a foundation for such a theoretical view, but more needs to
be done. Issues for future investigation include delineating the details of the rela-
tionship between cultural knowledge and embodiment, such as how the relation-
ship is managed in construal operations, both online and offline.
A related issue is dynamism and linguistic structure. Langacker (1994) has
argued that grammar transforms the meanings of words and phrases to create new
meanings; the structural components of language, such as grammar, phonology,
and morphology, may resist dynamic construal because of the tight relationship
between form and meaning. Based on the results of the current studies, semantic
meaning does show some effects of dynamic construal; this was seen in the data
samples of the blood and spleen metaphors that violated the structure of the con-
ceptualization and/or the extant cultural model. However, language structure may
not be so malleable. The current study indicates that grammar can be subject to
the effects of dynamic construal (or creative variations in linguistic structure) be-
cause only one data sample out of the total collected was found that followed the
prototypical form (i.e., His blood boiled). More research is needed to delineate the
specific effects of dynamic construal on the structural aspects of language in-
cluding phonology, morphology, and syntax.

Experiential scenes, domain matrices, and non-autonomous knowledge

In current cognitive-functional theory, the experiential scene and the domain ma-
trix are viewed as the fundamental constructs in the cognitive conceptualization
process. However, few studies employ these constructs to investigate conceptual
metaphor. The scene is a key component in Lakoff and Kvecses (1987) descrip-
tion of the Anger Prototype Scenario, and the domain matrix is a key concept in
Croft (2003/1993), but other studies that employ the experiential scene or the do-
main matrix to test various aspects of a conceptualization are relatively rare. The
current studies have shown that both concepts help to delineate important aspects
of non-autonomous knowledge, such as the role of cultural models to provide per-
spective on the scene, the relationships between concepts in the domain matrix of
emotion, and the prototypicality of a specific conceptualization. We recommend
further investigation of these important constructs in studies of conceptual meta-
phor and non-autonomous knowledge.

Cultural models and linguistic forms

As well, the relationship between a linguistic form and the cultural models that
provide perspective on the experiential scene for the speech community must be
studied in detail. The results of the studies in this volume show that the relationship
Chapter 9. The non-autonomous nature of cognition, language, and culture

between conceptualization and cultural models is an intimate one; the meaning

(and sometimes, the form) of a metaphorical expression was significantly influ-
enced by the cultural model(s) licensed by the speech community during a particu-
lar historical period. This relationship was seen throughout the five centuries of data
under study but was most clearly visible beginning in the latter half of the 19th cen-
tury, when the unified model was largely replaced by the cooking semantic frame.
This shift in the cultural model brought about a significant change in perspective
that in turn affected the structure and meaning of the data samples collected be-
tween 1850 and 1990. The non-autonomous nature of knowledge, encompassing
both the cognitive and cultural types, implies that the contemporary forms of lan-
guage in any time period will respond to shifts in embodied experience and shared
cultural knowledge. Language changes constantly in response to changes in human
experience of both the physical world and the cultural world of a speech commu-
nity. It is recommended that research in metaphor and cognitive-functionalist work
in general look carefully at cultural models at work in linguistic forms, in order to
describe that form in its situational use in more detail and more accurately.
In addition, future research should investigate further the prototypicality of
the blood and spleen metaphors. The current studies indicated that both were pro-
totypical historically, primarily due to their relationship to the unified model, but
there are important issues concerning the nature of their usage in discourse. For
example, it was suggested in Chapter 3 and Chapter 6 that the blood and spleen
metaphors were in complementary distribution; that is, each metaphor was used
in particular and separate discourse situations to communicate specific perspec-
tive on the experiential scene. The blood metaphor was employed to communicate
public, violent anger, and the spleen metaphor conversely used for private, non-
violent emotions, including anger. If this analysis is correct, then each conceptual-
ization was constrained in its pragmatic use, possibly due to the specific experien-
tial scene that motivated each conceptualization. While the results of the current
study suggest that the two concepts were in complementary distribution in lan-
guage use, further study is needed to investigate this question in more detail. Such
study would aid to corroborate the prototypicality of the two metaphors as well as
the pragmatic justifications for their use in real discourse.

Diachronic studies of conceptual metaphor

More diachronic research is also needed in conceptual metaphor, a point that has
been discussed throughout this volume. The review of the literature in historical
studies of CM showed that this line of research is of recent origin and the number
of studies to date is low compared to other aspects of conceptual metaphor re-
search. The current studies have demonstrated the usefulness of diachronic study,
Metaphor across Time and Conceptual Space

not only for understanding historical linguistic forms, but also for understanding
their current, synchronic descendants, as Bybee, Sweetser, and others have long
argued. The relationship between a contemporary linguistic expression and its his-
torical precursors is crucial for delineating the details of conceptualization and
cultural models of both forms.

The social context of language forms

Related to the study of metaphor in cultural and historical context of metaphor use
is parallel work in social context. Crofts (2008) concept of the semiotic triangle
can be applied to any time period and any linguistic form. The current studies have
demonstrated that understanding the shared common ground of the speaker/writ-
er and hearer/listener provides data that aids the analysis of the semantic meaning
of linguistic forms, and this study can be extended to the study of pragmatic use.
In cognitive-functional research, a promising new area of investigation intended
to study the relationship between linguistic form and its social context is Cognitive
Sociolinguistics. An edited volume of studies in this area was recently published
(Kristiansen & Dirven, 2008). This area of research would be a natural one for
studies that combine the cultural, historical, and social contexts of language use to
develop multi-level investigations that provide detailed descriptions of language
structure, meaning and use across speech communities and the time course.

Methodological considerations in cognitive-functional research

There are several issues in CF research methods that the current studies implicate.
These include (1) multidisciplinary research methods; (2) corpus size; (3) CADS
and mixed-methods research; (4) non-linguistic data; (5) additional research tech-
niques. Each is discussed in turn below.

Multidisciplinary research

As the issues discussed in Chapter 1, 2 and 3 have shown, investigations into con-
ceptualization and cultural models benefit from the use of multidisciplinary
methodologies, including the longitudinal study design, the application of non-
linguistic data, compiled corpora, and frequency statistics. The changes seen in the
metaphoric expressions and in the DM of emotion would not have been as visible
without these specific techniques. Specific to studies of conceptual metaphor, in-
vestigating a specific conceptualization requires analyses of multiple instances of
linguistic samples that hypothetically map the target domain. Analysis of individual
Chapter 9. The non-autonomous nature of cognition, language, and culture

samples, collected either via introspection or synchronic-historical study designs,

is insufficient by itself to confirm the target domain, and the results of such studies
cannot be generalized to reach conclusions concerning the synchronic or dia-
chronic prototypicality of a conceptualization. The research design employed in
the current studies illuminated some important ways in which conceptualizations
vary over time. The multidisciplinary nature of the current study design brought
out these aspects more clearly, and the results of the study have demonstrated the
advantages of the methodologies employed. It is recommended that future studies
of conceptual metaphor employ one or more of these research techniques.

Corpus size

In order to increase the usefulness of diachronic research, historical corpora need to

increase in size. The current study totaled 3.7 million words, which is large for dia-
chronic studies of language; however, currently available diachronic corpora are sig-
nificantly smaller than synchronic corpora. For example, the synchronic British
National Corpus (BNC) includes over 100 million words published between 1990
and the present, and the corpus continues to grow every year. In contrast, the largest
historical corpus available until 2011 (see below) comprised less than 5 million
words. Corpus size is particularly important for studies of metaphor because meta-
phor as a form has a relatively low frequency of use compared to other linguistic
forms, such as nouns, verbs, and prepositions (Biber, 2006). In addition, the ability
to generalize the results of corpus studies increases as corpus size increases, espe-
cially in lexical (i.e., keyword) studies (Biber et al., 1998, p. 30). Research in dia-
chronic language, both quantitative and qualitative, would benefit significantly from
larger total word sizes in historical corpora. The development of new, representative,
high word-count corpora is an important effort to improve historical research in
general and, in particular, for the study of low-frequency forms, such as metaphor.
In this regard, new, high word-count historical corpora are either in develop-
ment or have been recently released. These include the Corpus of Historical
American English (COHA), available for public use in 2011. The corpus includes
400 million words from written texts published between 1810 and 2009. The COHA
corpus is almost 100 times larger than any other historical corpus currently in ex-
istence. The texts selected for the corpus maintain a balance between academic,
literary, newspapers, and popular press magazine genres in each decade of the his-
torical period, in order to increase the representativeness of the corpus. The COHA
project was led by Mark Davies of Brigham Young University and funded by the
National Endowment for the Humanities. COHA opens a new era in diachronic
research in historical texts that will allow for empirical research in a corpus
that spans 200 years of texts within the U.S. speech community. COHA offers
Metaphor across Time and Conceptual Space

opportunities for increasing the power of research studies to understand language

within a historical period and language change over time, as well as the factors that
affect those changes, including conceptualization and cultural models.
In order to use the large word-count corpora, the current research method
would need revision, specifically in data sampling procedures. In the COHA cor-
pus, the number of instances of a keyword could number in the millions, and a
single metaphoric expression could top 1,000 cases. The CADS method of close
analysis of each instance in extended discourse means that analyzing every in-
stance of a specific metaphor would take an inordinate amount of time, effort, and
manpower. Considering the amount of information collected for the current stud-
ies, CADS should not be overlooked as a research method simply due to the em-
barrassment of riches available in large word-count corpora; scientific sampling
procedures, specifically random sampling, will be necessary to collect data samples
for study. As seen in the current studies, currently-available scientific research
practices can resolve the potential problems in research design.

CADS and mixed-methods research

The Corpus-Assisted Discourse Studies system is a mixed-methods research de-

sign, combining both quantitative and qualitative empirical data collection and
analysis techniques. Mixed designs have some useful advantages for the study of
conceptualization and culture, compared to designs that employ either qualitative
or quantitative techniques exclusively; the most significant results of the current
studies would not have been possible in a purely quantitative or qualitative study.
CADS also allowed the collection and analysis of non-linguistic data, such as the
paintings, graphics, and songs analyzed for the ancillary study. The data were use-
ful for detailing the cultural knowledge of the unified model among English speak-
ers in the historical period under study, supporting the arguments of Lucy (1996)
and Enfield (2002) that such data are important for the study of culture and lan-
guage. Yet, mixed designs are generally employed much less frequently than the
purely quantitative or qualitative designs. To be fair, there are important logistical
issues (primarily, time and researcher expertise) that may discourage the use of
mixed-methods designs; however, the multi-level nature of the data collected for
analysis can increase the power and significance of the results. We recommend
serious consideration of mixed methods designs for studies investigating research
questions that can take advantage of its unique characteristics.

Non-linguistic data

Another area of investigation in CF that is useful for the study of conceptualization

and culture is the use of non-linguistic data. In the current studies, the quantitative
Chapter 9. The non-autonomous nature of cognition, language, and culture

and qualitative analyses of the linguistic data were more straightforward and effi-
cient due in part to the non-linguistic data that was collected for the ancillary
study. Without the non-linguistic data, the categorization and analysis of the lin-
guistic samples collected in the main study of historical metaphor (see Table 3,
including the Normalized Frequency Rate or NFR) would have been more difficult
and time-consuming. In addition, because the contemporary researchers knowl-
edge of historical linguistic forms is not reliable for analyzing diachronic data, the
non-linguistic data collected in the ancillary study and employed in the main study
to analyze the diachronic metaphor samples increased the accuracy of the analysis.
Finally, the techniques developed in this research area can be employed to analyze
non-linguistic data of many types, such as graphic representations, musical com-
positions, and physical objects.
The new research field of multimodal metaphor employs non-linguistic data
as a standard research strategy. As was discussed in Chapter 1, the multimodal
metaphor data collected for the current studies primarily paintings, graphics,
and music were helpful for describing both conceptualization and cultural mod-
els of the metaphors found in the diachronic data. An edited volume of studies in
multimodal metaphor was recently published (Forceville & Urios-Aparisi, 2009).
Multimodal metaphor, like cognitive sociolinguistics, is a new, complementary re-
search area that can usefully inform the multi-level research conducted in the cur-
rent studies and future work in mixed-methods designs within cognitive-func-
tional research.

Additional research techniques

In addition to the qualitative (i.e., CADS) and quantitative methods (i.e., NFR)
employed in the current studies, it is recommended that future CF research con-
tinue to add a variety of techniques to analyze linguistic data. These techniques
include inferential statistics and advanced analysis metrics, all of which require
computer technology to aid in the categorization and analysis of data. Recent stud-
ies and edited volumes in CF that have employed statistical techniques and ad-
vanced technology, including conceptual metaphor study, include Ghesquire and
Vandevelde, 2011; Oster, 2010; Sim, 2011; and, Stefanowitsch and Gries, Eds.,
2006. These studies and others are notable for employing digital text corpora,
which can be easily manipulated by statistical measures.6

6. Quantitative statistical techniques do not preclude the use of qualitative analysis,; CADS
analysis and other discourse analysis techniques can be analyzed with quantitative measures, as
Partington (2004) argued (see Chapter 1 for discussion). In addition, qualitative techniques are
empirical in nature and therefore support scientifically-valid research.
Metaphor across Time and Conceptual Space

CF researchers can and should pursue the empirical study of data collected
from a large sample of native speakers to support the non-objectivist research en-
terprise. Current CF theory, established research techniques, and technological
innovations in text corpora and computer technology provide exciting new venues
and methods to investigate cognitive conceptualization via diachronic language
data. In light of these future possibilities, we look forward to new studies of meta-
phor across time and conceptual space.

Chapter summary

The chapter provided a discussion of the results of the three studies, including the
implications for the guiding principles of cognitive-functionalism, conceptual
metaphor theory, and future research. A brief synopsis of the results is presented
1. The blood and spleen conceptual metaphors constitute separate historical pro-
totypes due to their distinct experiential scenes, conceptual domains, and cog-
nitive mappings;
2. The two prototypes were part of a large, complex array of related concepts,
which included both embodied experience and cultural models, termed here
the Domain Matrix (DM) of emotion (or DME);
3. The DME conceptualized anger and a variety of different emotion con-
4. The DME also had several important dimensions, constituted as scales, in-
cluding temperature, reason, and control, which were used to organize
the emotion concepts into a complex system of relations;
5. A speaker/writer could employ the DME to combine aspects of embodied ex-
perience with principles of the unified model and create a unique meaning to
meet the communicative needs of the discourse situation;
6. The conceptualizations of emotion in the blood and spleen metaphors varied
in structure and frequency over time, indicating that the metaphors were in-
fluenced by the historical unified model of human physiology;
7. Changes in cultural knowledge over time affected the frequency of use of a
language form. This result was found in two different ways: (1) the increase
and decrease in the popularity of the unified model also coincided with the
increase and decrease in the use of the blood and spleen metaphors; (2) the
frequency and structure of the blood metaphor samples employing the boil-
keyword changed when the unified model was replaced with the cooking
semantic frame;
Chapter 9. The non-autonomous nature of cognition, language, and culture

8. Implications for cognitive-functionalism and conceptual metaphor theories

include the inseparable nature of embodiment and cultural models; the con-
struct of dynamic construal; the false dichotomy of the embodied core vs. the
cultural periphery; the contributions of the constructs of the experiential
scene, the domain matrix, and non-autonomous knowledge to conceptual
metaphor theory; the usefulness of frequency statistics for investigating proto-
typicality; and, the complex relationship between synchronic forms and their
diachronic precursors;
9. Future research includes the areas listed in #8 above plus the methodological
implications of mixed-methods research, the usefulness of corpus studies for
conceptual metaphor research, the importance of non-linguistic data in stud-
ies of conceptualization and culture, due to the inaccuracy of contemporary
linguistic competence to analyze accurately the semantic meaning of histori-
cal language forms, and the need to add other research techniques, not cur-
rently in use in cognitive-functional and conceptual metaphor research, that
may aid the process of scientific discovery.
The studies described in this volume indicate that cultural knowledge as constitut-
ed in systems of cultural models influence conceptualization at the level of the con-
ceptualization process itself. This conclusion supports the work of other researchers,
including Cienki (1999); Croft (1993; 2002; 2008); Emanatian (1999); Enfield
(2002); Geeraerts and Gevaert (2008); Geeraerts, Gevaert, and Speelman (2011);
Gevaert (2002); Gibbs (1999); Goldberg (1995; 2006; 2010); Kovecses (2005; 2008;
2009; 2010a); Kovisto-Alanko and Tissari (2006); Niemeier (2008); Sim (2011);
Sinha and Jensen de Lpezj (2000); Trim (2011); and, Yu (2009). Therefore, these
studies constitute a body of work that demonstrates the important role that cul-
tural knowledge plays in interpreting the experiential scene within a speech
community to produce linguistic expressions that can be comprehended by the
members of the community via shared common ground. As with other areas of
cognitive-functional research, the major conclusion is the same for all the studies
shown above: the purpose of general cognitive processes is the construal of the
physical world for communicating that experience within a particular speech com-
munity. With such real world goals in mind, the Epilogue describes briefly an ap-
plication of the current studies to the teaching and learning of metaphor in the
second language classroom.

Bridging the Gap between theory

and real-world language use


The conclusions discussed in the previous chapter have important implications for
language education, particularly second and foreign languages. The conceptual
unity of cognition and cultural models in frames suggests that teaching and study-
ing a specific linguistic metaphor requires that teachers and students account for
both aspects, in order to understand fully the form, meaning, and pragmatic use
of the metaphor for the purpose of communicating in speaking and writing. In
addition, it is assumed here that the learning of metaphor (and its related form,
metonymy) is necessary for second language students to communicate effectively
in the target language, as previous research has found (e.g., Barcelona, 2010;
Berndi, Csbi, & Kvecses, 2008; Cameron, 2003; Cameron & Deignan, 2006;
Littlemore, 2003, 2009; Littlemore & Low, 2006; MacArthur, 2010). The cognitive
and cultural aspects of conceptual metaphors discussed in this volume indicate
that cognitive conceptualizations play important roles in language learning and
teaching. The goal of this Epilogue is to discuss briefly some justifications for ap-
plying the results of the studies in this volume to a contemporary, real-world prob-
lem: the teaching of metaphor to second language learners.

The usage-based model and language learning

A fundamental principle from the Usage-based Model of Language mentions the

effects that language use have on language acquisition and learning (Barlow &
Kemmer, 2000, p. xi).
On this issue, Barlow and Kemmer argue that performance is a part of the
speakers competence; they are not separate aspects of language, as generative
grammar theory contends. In addition, knowledge of language is created by lan-
guage use, as Barlow and Kemmer explain:
Metaphor across Time and Conceptual Space

Since in a usage-based model instances of producing and understanding language

are of central importance to the structuring of the linguistic system, they must
be especially significant in the acquisition of language, when the system is in the
process of taking form (2000, p. xi).

Language use promotes language learning because each usage event (see Chapter 1)
adds to and revises the learners knowledge of the linguistic system and its schemas
the system increases in its detailed knowledge of form and meaning through
experience with the language. As well, the use of the linguistic system within a
particular speech community provides knowledge of the pragmatic uses of lan-
guage forms and meanings. In this way, language use supports and increases
knowledge in all three parts of non-autonomous knowledge described by the se-
miotic triangle (Croft, 2008).
Therefore, increasing the learners knowledge and use of a language form in the
second language requires increasing the learners experience with the form and
with its use in real-world communication. This is not a controversial idea, as many
pedagogical systems assume that experience is important to learning, yet often this
principle is not enacted in tangible ways in the classroom. Research results can and
should be employed to aid second language learners to make better use of the ex-
perience with language to increase their knowledge (and in turn their future use) of
the language. This effort is in line with both the theoretical principles of the Usage-
based Model and the goals of second language teaching practice. The overall objec-
tive, therefore, is to find ways to bridge the gap between theory and practice to
find solutions to real world problems. Language learning and teaching, especially
for students learning a second language, is a natural place to center such an effort.

Cognitive-functional research in language teaching methods

Cognitive-functional researchers have been working to apply CF research to de-

velop practical teaching methods over the past 20 years or so; several of the studies
related to metaphor and metonymy are cited at the beginning of this chapter. Ad-
ditional examples include Grundy (2004), Langacker (2001), Littlemore (2003),
and Tyler and Evans (2001), among others. Several edited collections of studies
have been published, including Ptz, Niemeier, and Dirven (2001), Achard and
Niemeier (2004), Sharifian and Palmer (2007), and Boers and Lindstromberg
(2008). Most recently, in 2010 a special issue of AILA Review, edited by Littlemore
and Juchem-Grundmann, was devoted to the topic, Applied cognitive linguistics
in second language learning and teaching, including the Barcelona and MacArthur
studies cited previously. These studies and others in the teaching and learning
of metaphor and metonymy have been discussed at a variety of conferences,

including the annual Conceptual Structure, Discourse, and Language (CSDL) and
the International Cognitive Linguistics Association meetings. Finally, several the-
orists and researchers have described general research goals and study programs
in second language learning and teaching from a CF point of view; for examples,
see Dirven (1989) and Taylor (1993), as well as the recommendations presented in
Langacker (2001) and Tyler and Evans (2001). The work done so far has delineated
a number of concepts and issues that must be accounted for when applying CF
theory to the teaching and learning of metaphor in the second language class-
room. In addition, the extant studies have indicated that CF has the potential to
aid and enhance classroom pedagogical practices. It is recommended that these
efforts continue and are increased and expanded.

Future research

The role of cultural models

From the point of view of the studies presented in this volume, future work should
include several issues brought out by the study results (see Chapter 9). First, a great-
er emphasis on the effects of cultural models on metaphor comprehension, produc-
tion, and pragmatic use is warranted. The interplay between embodied experience
and cultural models is complex, and this relationship will also affect the learning and
teaching of metaphor and metonymy. A few studies have been done (see Littlemore,
2003; 2009, for example), but more work is clearly needed to understand this rela-
tionship in more detail, in order to devise effective teaching practices.

The role of language use

Second, studies of language use are needed, especially corpus studies. Corpus
studies in cognitive-functional linguistics have increased greatly over the past de-
cade as researchers have come to understand the value of studying language use in
order to delineate the conceptualizations that underlie linguistic expressions.
However, corpus study is still in the early stages in CF. More studies and more
detailed analyses are needed in order to understand language use and its relation-
ship to cognition.

The role of diachronic metaphor

Finally, more diachronic studies of metaphor are needed. This volume has reiter-
ated and demonstrated the value of a fundamental principle, that the relationship
Metaphor across Time and Conceptual Space

between synchronic forms and their diachronic ancestors is important and affects
contemporary language knowledge and its use in communication. Research to un-
derstand this relationship in detail will, in the long run, help to reveal important
principles about language, its use, and ultimately, conceptualization. The current
studies have indicated the important influence that cultural models exert on both
linguistic expressions and cognitive conceptualization. To understand contempo-
rary language, we must look to the past, as well as the present.


To conclude, cognitive-functionalism is a theory of human experience in the world

that can have real-world, practical implications. Any theory that purports to de-
scribe the world as we know will have the potential to help solve human problems.
Researchers can and should include this important research objective in their re-
search designs. A natural place to pursue this objective is the study of language
learning, including teaching methods. Three areas of study suggested for investi-
gating language learning within a CF framework include cultural models, language
use, and diachronic metaphor.

The historical Four Humors texts

with brief annotations

Barrough, P. (1590). The method of phisick. London: Richard Field.

Electronic facsimile (PDF) of 1590 edition; 2nd of ten editions.
Boorde, A. (1542). Dyetary of helth. London: Robert Wyer.
Electronic facsimile (PDF) of 1542 edition; 1st of five editions.
Bright, T. (1613). A treatise of melancholy. London: William Stansby.
Electronic facsimile (PDF) of 1613 edition (2 editions in that year); 3rd of four editions.
Burton, R. (1621). The anatomy of melancholy. Oxford: John Lichfield and James Short.
Printed edition of the 1621 text; 1st of nine editions in the 17th century.
Charron, P. (1630). Of wisdome. London: George Miller.
Electronic facsimile (PDF) of 1630 edition; 4th of nine editions.
Coffeteau, N. (1621). A table of humane passions (E. Grimeston, Trans.). London: Nicholas
Electronic facsimile (PDF) of 1621 edition; 1st of one edition.
Cogan, T. (1605). The hauen of health. London: Melch. Bradwood.
Electronic facsimile (PDF) of 1605 edition; 5th of seven editions.
Cuff, H. (1640). The differences of the ages of mans life. London: Thomas Harper.
Electronic facsimile (PDF) of the 1640 edition; 3rd of three editions.
Dariot, C. (1598). The astrologicall iudgement of the starres. London: Thomas Purfoot.
Electronic facsimile (PDF) of the 1598 edition; 3rd of three editions.
de Glanville, B. (1582). De proprietatibus rerum (J. Trevisa, Trans.). London: Thomas East.
Electronic facsimile (PDF) of 1582 reprint; English translation of Latin text written in 1360. 4th
of four editions.
de Mediolano, J. (1609). The Englishmans doctor. London: S. Stafford.
Electronic facsimile (PDF) of the 1609 edition; 3rd of five editions.
Elyot, T. S. (1610). The castle of health. London: W. Jaggard.
Electronic facsimile (PDF) of the 1610 edition; 16th of 16 editions.
Huarte, J. (1698). Examen de ingenios: Or, the tryal of wits (M. Bellamy, Trans.). London: Richard
Electronic facsimile (PDF) of 1698 edition; 7th of seven editions.
Lemnius, L. (1581). The touchstone of complexions (T. Newton, Trans.). London: Thomas
Electronic facsimile (PDF) of 1581 edition; 3rd of five editions.
Moulton, T. (1546). Myrrour or glasse of helth. London: Author.
Electronic facsimile (PDF) of 1546 edition; 8th edition of 14 editions.
Rogers, T., & H, W. (1580). A paterne of a passionate minde. London: Thomas East.
Electronic facsimile (PDF) of the 1580 edition; 2nd of two editions of an abridged version of the
1576 text.
Metaphor across Time and Conceptual Space

Walkington, T. (1607). The optick glasse of humors. London: John Windet.

Electronic facsimile (PDF) of the 1607 edition; 1st of four editions.
Wright, T. (1601). The passions of the minde. London: Valentine Simmes.
Electronic facsimile (PDF) of the 1601 edition; 2nd of six editions.


Achard,M., & Niemeier,S. (Eds.). (2004). Cognitive linguistics, second language acquisition, and
foreign language teaching. Berlin: Mouton de Gruyter.
Ackerknecht,E. H. (1982). A short history of medicine (Rev. ed.). Baltimore: Johns Hopkins Uni-
versity Press.
Allan,K. (2008). Metaphor and metonymy: A diachronic approach. Malden, MA: Wiley-Blackwell.
Babb,L. (1951). The Elizabethan malady: A study of Melancholia in English literature from 1580
to 1642. East Lansing, MI: Michigan State College Press.
Barcelona,A. (2010). Metonymic inferencing and second language acquisition. AILA Review,
23, 134154.
Barcelona,A., & Soriano,C. (2004). Metaphorical conceptualization in English and Spanish.
European Journal of English Studies, 8(3), 295307.
Barlow,M., & Kemmer,S. (2000). Introduction: A usage-based model of language. In M. Barlow
and S. Kemmer (Eds.), Usage based models of language (pp. viixxviii). Stanford, CA: CSLI.
Beck,A. H. (2004). The Flexner Report and the standardization of American medical education.
Medical Student JAMA, 291(17), 21392140.
Berndi,M., Csbi, S., & Kvecses,Z. (2008). Using conceptual metaphors and metonymies in
vocabulary teaching. In F. Boers & S. Lindstromberg (Eds.), Cognitive linguistic approaches
to teaching vocabulary and phraseology (pp. 6599). Berlin: Mouton de Gruyter.
Bertuol,R. (2001). The square circle of Margaret Cavendish: The 17th-century conceptualiza-
tion of mind by means of mathematics. Language and Literature, 10(1), 2139.
Biber,D. (2006). University language: A corpus-based study of spoken and written registers.
Amsterdam: John Benjamins.
Biber,D., Conrad, S., & Reppen,R. (1998). Corpus linguistics: Investigating language structure
and use. Cambridge: Cambridge University Press.
Biber,D., Finegan,E., & Atkinson,D. (1994). ARCHER and its challenges: Compiling and ex-
ploring A Representative Corpus of Historical English Registers. In U. Fries, P. Schneider,
& G. Tottie (Eds.), Creating and using English language corpora. Papers from the 14th Inter-
national Conference on English Language Research on Computerized Corpora, Zurich 1993
(pp. 113). Amsterdam: Rodopi.
Boers,F. (1999). When a bodily source domain becomes prominent. In R. W. Gibbs and G. J.
Steen (Eds.), Metaphor in Cognitive Linguistics (pp. 4756). Amsterdam: John Benjamins.
Boers,F., & Lindstromberg,S. (Eds.). (2008). Cognitive linguistic approaches to teaching vocabu-
lary and phraseology. Berlin: Mouton de Gruyter.
Boorde,A. (1542). Dyetary of helth. London: Robert Wyer.
Bruce,L. (1988). Serialization: From syntax to lexicon. Studies in Language, 12(1), 1949.
Burton,R. (1932/1621). The anatomy of melancholy. Oxford: John Lichfield and James Short.
Bybee,J. (1988). The diachronic dimension in explanation. In J. A. Hawkins (Ed.), Explaining
language universals (pp. 350379). Malden, MA: Blackwell.

Bybee,J. (2001). Main clauses are innovative, subordinate clauses are conservative. In J. Bybee &
M. Noonan (Eds.), Complex sentences in grammar and discourse: Essays in honor of Sandra
A. Thompson (pp. 117). Amsterdam: John Benjamins.
Bybee,J. (2003). Mechanisms of change in grammaticization: The role of frequency. In B. D.
Joseph & R. D. Janda (Eds.), Handbook of historical linguistics (pp. 602623). Malden, MA:
Bybee,J. (2010). Language, usage and cognition. Cambridge: Cambridge University Press.
Bybee,J., Perkins,R., & Pagliuca,W. (1994). The evolution of grammar: Tense, aspect, and modal-
ity in the languages of the world. Chicago: University of Chicago Press.
Caballero,R. (2003). Metaphor and genre: The presence and role of metaphor in the building
review. Applied Linguistics, 24, 145167.
Cameron,L. (2003). Metaphor in educational discourse. London: Continuum.
Cameron,L. (2008). A discourse approach to metaphor: Explaining systematic metaphors for
literacy processes in a school discourse community. In A. Tyler, Y. Kim, & M. Takada (Eds.),
Language in the context of use: Usage-based approaches to language and language learning
(pp. 321339). New York: Mouton de Gruyter.
Cameron,L., & Deignan,A. (2006). The emergence of metaphor in discourse. Applied Linguis-
tics, 27(4), 671690.
Casasanto,D. (2009). When is a linguistic metaphor a conceptual metaphor? In V. Evans & S.
Pourcel (Eds.), New directions in cognitive linguistics (pp. 127145). Amsterdam: John
Charteris-Black,J. (2003). Speaking with a forked tongue: A comparative study of metaphor and
metonymy in English and Malay phraseology. Metaphor and Symbol, 18, 289310.
Chomsky,N. (1957). Syntactic structures. The Hague: Mouton.
Chomsky,N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Chomsky,N. (1975). Reflections on language. New York: Pantheon.
Cienki,A. (1998). Metaphoric gestures and some of their relations to verbal metaphoric
expressions. In J.-P. Koenig (Ed.), Discourse and cognition: Bridging the gap (pp. 189204).
Stanford, CA: CSLI.
Cienki,A. (1999). Metaphors and cultural models as profiles and bases. In R. W. Gibbs, Jr. & G.
J. Steen (Eds.), Metaphor in cognitive linguistics: Selected papers from the fifth International
Cognitive Linguistics Conference, Amsterdam, July, 1997 (pp. 189203). Amsterdam: John
Clark,H. H. (1996). Communities, commonalities, and communication. In J. Gumperz & S. C.
Levinson (Eds.), Rethinking linguistic relativity (pp. 324355). Cambridge: Cambridge Uni-
versity Press.
Croft,W. (1998). Linguistic evidence and mental representations. Cognitive Linguistics, 9(2),
Croft,W. (2003/1993). The role of domains in the interpretation of metaphors and metonymies.
In R. Dirven & R. Prings (Eds.), Metaphor and metonymy in comparison and contrast
(pp. 161205). Berlin: Mouton de Gruyter.
Croft,W. (2008). Toward a social cognitive linguistics. In V. Evans & S. Pourcel (Eds.), New Di-
rections in Cognitive Linguistics (pp. 395419). Amsterdam: John Benjamins.
Croft,W. (2009). Connecting frames and constructions: A case study of eat and feed. Construc-
tions and Frames, 1(1), 728.
Croft,W., & Cruse,D. A. (2004). Cognitive linguistics. Cambridge: Cambridge University Press.
Metaphor across Time and Conceptual Space

Csbi,S. (2001). The concept of America in the Puritan mind. Language and Literature, 10(3),
Cuff,H. (1640). The differences of the ages of mans life. London: Thomas Harper.
DAndrade,R. (1987). A folk model of the mind. In D. Holland & N. Quinn, Cultural models in
language and thought (pp. 112148). New York: Cambridge University Press.
Dariot,C. (1598). The astrologicall iudgement of the starres. London: Thomas Purfoot.
de Mediolano,J. (1609). The Englishmans doctor. London: S. Stafford.
Deignan,A. (2003). Metaphorical expression and culture: An indirect link. Metaphor and Sym-
bol, 18, 255271.
Deignan,A. (2005). Metaphor and corpus linguistics. Amsterdam: John Benjamins.
Deignan,A. (2006). The grammar of linguistic metaphors. In A. Stefanowitsch & S. Th. Gries
(Eds.), Corpus-based approaches to metaphor and metonymy (pp. 106122). Berlin: Mouton
de Gruyter.
Deignan,A., & Potter, L. (2004). A corpus study of metaphors and metonyms in English and
Italian. Journal of Pragmatics, 36, 12311252.
Diller,H.-J. (1994). Emotions in the English lexicon: A historical study of the lexical field. In F.
M. Fernndez, M. Fuster, & J. J. Calvo (Eds.), English Historical Linguistics 1992, 219234.
Amsterdam: John Benjamins.
Dirven,R. (1989). Cognitive Linguistics and pedagogic grammar. In G. Leitner & G. Graustein
(Eds.), Linguistic theorizing and grammar writing (pp. 5675). Tbingen, Germany: Max
Domaradzki,M. (2011). The self in Arabic and the relativism-universalism controversy.
Cognitive Linguistics, 22(3), 535567.
Draper,J. W. (1945). The humors and Shakespeares characters. Durham, NC: Duke University
Durie,M. (1997). Grammatical structures in verb serialization. In A. Alsina, J. Bresnan, & P.
Sells (Eds.), Complex predicates (pp. 289354). Stanford, CA: CSLI.
Eerden,B. (2009). Anger in Asterix: The metaphorical representation of anger in comics and
animated films. In C. J. Forceville & E. Urios-Aparisi, Multimodal metaphor (pp. 243264).
Berlin: Mouton de Gruyter.
Emanatian,M. (1999). Congruence by degree: On the relation between metaphor and cultural
models. In R. W. Gibbs, Jr. & G. J. Steen (Eds.), Metaphor in cognitive linguistics: Selected
papers from the fifth International Cognitive Linguistics Conference, Amsterdam, July, 1997
(pp. 205218). Amsterdam: John Benjamins.
Enfield,N. J. (2000). On linguocentrism. In M. Ptz & M. Verspoor (Eds.), Explorations in lin-
guistic relativity (pp. 125157). Amsterdam: John Benjamins.
Enfield,N. J. (2002). Cultural logic and syntactic productivity: Associated posture constructions
in Lao. In N. J. Enfield (Ed.), Ethnosyntax: Explorations in grammar and culture (pp. 231258).
Oxford: Oxford University Press.
Estes,J. W., & Goodman,D. M. (1986). The changing humors of Portsmouth: The medical biogra-
phy of an American town, 16231983. Boston: F.A. Countway.
Everett,D. (2005). Cultural constraints on grammar and cognition in Pirah: Another look at
the design features of human language. Current Anthropology, 46(4), 621634.
Fauconnier,G., & Turner,M. (2003). The way we think: Conceptual blending and the minds hid-
den complexities. New York: Basic Books.
Fesmire,S. A. (1994). What is cognitive about Cognitive Linguistics? Metaphor and Symbolic
Activity, 9(2), 149154.

Fillmore,C. J. (1982). Frame semantics. In Linguistics in the morning calm: Selected paper from
Sicol 1981 (pp. 111137). Seoul, Korea: Hanshin.
Flexner,A. (1910). Medical education in the United States and Canada. New York, NY: Carnegie
Foundation for the Advancement of Teaching.
Forceville,C. J., & Urios-Aparisi,E. (Eds.). (2009). Multimodal metaphor. Berlin: Mouton de
Geeraerts,D. (2006). Introduction: A rough guide to Cognitive Linguistics. In D. Geeraerts
(Ed.), Cognitive linguistics: Basic readings (pp. 128). Berlin: Mouton de Gruyter.
Geeraerts,D. (2010). Theories of lexical semantics. Oxford: Oxford University Press.
Geeraerts,D., & Gevaert,C. (2008). Hearts and (angry) minds in Old English. In F. Sharifian, R.
Dirven, N. Yu, & S. Niemeier (Eds.), Culture, body, and language: Conceptualizations of in-
ternal body organs across cultures and languages (pp. 319347). Berlin: Mouton de Gruyter.
Geeraerts,D., Gevaert, C., & Speelman,D. (2011). How anger arose: Hypothesis testing in dia-
chronic semantics. In K. Allan & J. A. Robinson (Eds.), Current methods in historical se-
mantics (pp. 109132). Berlin: Mouton de Gruyter.
Geeraerts,D., & Grondelaers,S. (1995). Looking back at anger: Cultural traditions and meta-
phorical patterns. In J. R. Taylor & R. E. MacLaury (Eds.), Language and the cognitive con-
strual of the world (pp. 153179). Berlin: Mouton de Gruyter.
Gentner,D., & Gentner,D. R. (1983). Flowing waters or teaming crowds: Mental models of elec-
tricity. In D. Gentner & A. L. Stevens (Eds.), Mental models (pp. 99129). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Gevaert,C. (2002). The evolution of the lexical and conceptual field of ANGER in Old and
Middle English. In J. E. Diaz Vera (Ed.), A changing world of words: Studies in English his-
torical lexicography, lexicology and semantics (pp. 275299). Amsterdam: Rodopi.
Ghesquire,L., & Vandevelde,F. (2011). A corpus-based account of the development of English
such and Dutch zulk: Identification, intensification, and (inter)subjectification. Cognitive
Linguistics, 22(4), 765797.
Gibbs,R. W., Jr. (1999). Taking metaphor out of our heads and putting it in the cultural world.
In R. W. Gibbs, Jr. & G. J. Steen (Eds.), Metaphor in cognitive linguistics (pp. 145166). Am-
sterdam: John Benjamins.
Gibbs,R. W., Jr. (2006). Introspection and cognitive linguistics: Should we trust our own intu-
itions? Annual Review of Cognitive Linguistics, 4, 135151.
Goldberg,A. E. (1995). Constructions: A construction grammar approach to argument structure.
Chicago: University of Chicago Press.
Goldberg,A. E. (1998). Patterns of experience in patterns of language. In M. Tomasello (Ed.),
The new psychology of language: Cognitive and functional approaches to language structure,
Vol. 1 (pp. 203219). Mahwah, NJ: Erlbaum.
Goldberg,A. E. (2006). Constructions at work: The nature of generalization in language. Oxford:
Oxford University Press.
Goldberg,A. E. (2010). Verbs, constructions, and semantic frames. In M. Rappaport Hovav, E.
Doron, & I. Sichel (Eds.). Lexical semantics, syntax, and event structure (pp. 3958). Oxford:
Oxford University Press.
Goldwasser,O. (2005). Where is metaphor? Conceptual metaphor and alternative classification
in the hieroglyphic script. Metaphor and Symbol, 20(2), 95113.
Grace,G. (1987). The linguistic construction of reality. London: Croon Helm.
Grady,J. (1997). Theories are buildings revisited. Cognitive Linguistics, 8, 267290.
Metaphor across Time and Conceptual Space

Grady,J. (2008). Superschemas and the grammar of metaphorical mappings. In A. Tyler, Y.

Kim, & M. Takada. (Eds.), Language in the context of use: Discourse and cognitive approach-
es to language (pp. 339360). Berlin: Mouton de Gruyter.
Green,G. M. (1995). Ambiguity resolution and discourse interpretation. In K. van Deemter &
S. Peters (Eds.), Semantic ambiguity and underspecification (pp. 126). Stanford, CA: CSLI.
Green,M. (1936/1737). The spleen. London: Methuen.
Gries,S. Th., Hampe, B., & Schnefeld, D. (2005). Converging evidence: Bringing together ex-
perimental and corpus data on the association of verbs and constructions. Cognitive Lin-
guistics, 16(4), 635676.
Grondelaers,S., Geeraerts, D., & Speelman, D. (2006). A case for a cognitive corpus linguistics.
In M. Gonzalez-Marquez, I. Mittelberg, S. Coulson, & M. J. Spivey (Eds.), Methods in Cog-
nitive Linguistics (pp.149169). Amsterdam: John Benjamins.
Grundy,P. (2004). The figure / ground gestalt and language teaching methodology. In M. Achard
& S. Niemeier (Eds.), Cognitive linguistics, second language acquisition, and foreign language
teaching (pp. 119142). Berlin: Mouton de Gruyter.
Harvey,W. (1958/1628). Anatomical studies on the motion of the heart and blood (C. D. Leake,
Trans.). Springfield, IL: Charles C. Thomas.
Heine,B. (1997). Cognitive foundations of grammar. Oxford: Oxford University Press.
Heylen,K., Tummers,J., & Geeraerts, D. (2008). Methodological issues in corpus-based Cogni-
tive Linguistics. In G. Kristiansen & R. Dirven, Cognitive Sociolinguistics: Language varia-
tion, cultural models, and social systems (pp. 91128). Berlin: Mouton de Gruyter
Hiatt,M. D., & Stockton,C. G. (2003). The impact of the Flexner Report on the fate of medical
schools in North America after 1909 [Electronic Version]. Journal of American Physicians
and Surgeons, 8(2), 3740. Retrieved May 26, 2013, from
Huarte,J. (1698). Examen de ingenios: Or, the tryal of wits (M. Bellamy, Trans.). London: Richard
Hunston,S. (2002). Corpora in applied linguistics. Cambridge: Cambridge University Press.
Jackson,W. (1770). Collection of glees, catches, rounds. Unpublished manuscript.
Johnson,M. (1987). The body in the mind: The bodily basis of meaning, imagination, and reason.
Chicago: University of Chicago Press.
Keesing,R. M. (1979). Linguistic knowledge and cultural knowledge: Some doubts and specula-
tions. American Anthropologist, 81(1), 1436.
Klibansky,R., Panofsky,E., & Saxl,F. (1964). Saturn and melancholy: Studies in the history of
natural philosophy, religion, and art. London: Nelson.
Koivisto-Alanko,P., & Tissari,H. (2006). Sense and sensibility: Rational thought versus emo-
tion in metaphorical language. In A. Stefanowitsch & S. Th. Gries (Eds.), Corpus-based ap-
proaches to metaphor and metonymy (pp. 191213). Berlin: Mouton de Gruyter.
Kvecses,Z. (1991). Happiness: A definitional effort. Metaphor and Symbolic Activity, 6(1), 2946.
Kvecses,Z. (1995). Anger: Its language, conceptualization, and physiology in light of cross-
cultural evidence. In J. R. Taylor & R. E. MacLaury (Eds.), Language and the cognitive con-
strual of the world (pp. 181196). Berlin: Mouton de Gruyter.
Kvecses,Z. (2000). Metaphor and emotion: Language, culture, and body in human feeling.
Cambridge: Cambridge University Press.
Kvecses,Z. (2005). Metaphor in culture: Universality and variation. Cambridge: Cambridge
University Press.

Kvecses,Z. (2006). Language, culture, and mind: A practical introduction. Oxford: Oxford
University Press.
Kvecses,Z. (2008). On metaphors for emotion: A reply to Ayako Omori (2008). Metaphor and
Symbol, 23, 200203.
Kvecses,Z. (2009). Metaphor, culture, and discourse: The pressure of coherence. In A. Mulsolff
& J. Zinken (Eds.), Metaphor and discourse (pp. 1124). London: Palgrave Macmillan.
Kvecses,Z. (2010a). A new look at metaphorical creativity in cognitive linguistics. Cognitive
Linguistics, 21(4), 663697.
Kvecses,Z. (2010b). Metaphor: A practical introduction (2nd ed.). Oxford: Oxford University
Kvecses,Z. (2011). Methodological issues in conceptual metaphor theory. In S. Handl & H.-J.
Schmid (Eds.), Windows to the mind: Metaphor, metonymy and conceptual blending
(pp. 2339). Berlin: Mouton de Gruyter.
Kristiansen,G., & Dirven,R. (Eds.). (2008). Cognitive Sociolinguistics: Language variation, cul-
tural models, social systems. Berlin: Mouton de Gruyter.
Lakoff,G. (1986). A figure of thought. Metaphor and Symbolic Activity, 1(3), 215225.
Lakoff,G. (1987). Women, fire, and dangerous things. Chicago: University of Chicago Press.
Lakoff,G. (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor and
thought (pp. 202251). Cambridge: Cambridge University Press.
Lakoff,G., & Johnson,M. (1980). Metaphors we live by. Chicago: University of Chicago Press.
Lakoff,G., & Johnson,M. (1999). Philosophy in the flesh: The embodied mind and its challenge to
Western thought. New York: Basic Books.
Lakoff,G., & Kvecses,Z. (1987). The cognitive model of anger inherent in American English.
In D. Holland & N. Quinn (Eds.), Cultural models in language and thought (pp. 195221).
Cambridge: Cambridge University Press.
Lakoff,G., & Turner,M. (1989). More than cool reason: A field guide to poetic metaphor. Chicago:
University of Chicago Press.
Langacker,R. W. (1987). Foundations of cognitive grammar: Theoretical prerequisites (Vol. I).
Stanford, CA: Stanford University Press.
Langacker,R. W. (1988). A usage-based model. In B. Rudzka-Ostyn (Ed.), Topics in Cognitive
Linguistics (pp. 127161). Amsterdam: John Benjamins.
Langacker,R. W. (1994). Culture, cognition, and grammar. In M. Ptz (Ed.), Language contact
and language conflict (pp. 2553). Amsterdam: John Benjamins.
Langacker,R. W. (2001). Cognitive linguistics, language pedagogy, and the English present
tense. In M. Ptz, S. Niemeier, & R. Dirven (Eds.), Applied cognitive linguistics I: Theory and
language acquisition (pp. 339). Berlin: Mouton de Gruyter.
Langacker,R. W. (2002). A study in unified diversity: English and Mixtec locatives. In N. J.
Enfield (Ed.), Ethnosyntax: Explorations in grammar and culture (pp. 138161). Oxford:
Oxford University Press.
Littlemore,J. (2003). The effect of cultural background on metaphor interpretation. Metaphor
and Symbol, 18(4), 273288.
Littlemore,J. (2009). Applying Cognitive Linguistics to Second Language Learning and Teaching.
Basingstoke, UK: Palgrave Macmillan.
Littlemore,J., & Juchem-Grundmann,C. (Eds.). (2010). Applied cognitive linguistics in second
language learning and teaching. AILA Review, 23.
Littlemore,J., & Low,G. (2006). Figurative thinking and foreign language learning. Basingstoke,
UK: Palgrave MacMillan.
Metaphor across Time and Conceptual Space

Lucy,J. A. (1996). The scope of linguistic relativity: An analysis and review of empirical research.
In J. Gumperz & S. C. Levinson (Eds.), Rethinking linguistic relativity (pp. 3769).
Cambridge: Cambridge University Press.
Ludmerer,K. M. (1985). Learning to heal. New York: Basic Books.
Lutz,C. (1988). Unnatural emotions: Everyday sentiments on a Micronesian atoll and their chal-
lenge to Western theory. Chicago: University of Chicago Press.
Maalej,Z. (2004). Figurative language in anger expressions in Tunisian Arabic: An extended
view of embodiment. Metaphor and Symbol, 19(1), 5175.
MacArthur,F. (2005). The competent horseman in a horseless world: Observations on a conven-
tional metaphor in Spanish and English. Metaphor and Symbol, 20(1), 7194.
MacArthur,F. (2010). Metaphorical competence in EFL: Where are we and where should we be
going? A view from the language classroom. AILA Review, 23, 155173.
Matsuki,K. (1995). Metaphors of anger in Japanese. In J. R. Taylor & R. E. MacLaury (Eds.), Lan-
guage and the cognitive construal of the world (pp. 137151). Berlin: Mouton de Gruyter.
Mischler,J. J., III. (2008). A time for anger: Conceptions of human feeling in modern English, A.D.
15001990. (Doctoral dissertation). Available fromProQuest Dissertations and Theses da-
tabase. (UMI No. 3306990)
Moder,C. L. (2004). Ice box moms and hockey dads: Context and the mapping of N-N meta-
phorical expressions. In M. Achard & S. Kemmer (Eds.), Language, culture, and mind
(pp. 109121). Stanford, CA: CSLI.
Moder,C. L. (2008). Its like making a soup: Metaphors and similes in spoken news discourse.
In A. Tyler, Y. Kim, & M. Takada (Eds.), Language in the Context of Use: Discourse and
Cognitive Approaches to Language (pp. 301321). Berlin: Mouton de Gruyter.
Morgagni,G. (1960/1761). The seats and causes of diseases investigated by anatomy. New York:
Mller,C. (2008). Metaphors dead and alive, sleeping and waking: A dynamic view. Chicago:
University of Chicago Press.
Niemeier,S. (2008). To be in control: Kind-hearted and cool-headed. The head-heart dichotomy
in English. In F. Sharifian, R. Dirven, N. Yu, & S. Niemeier (Eds.), Culture, body, and lan-
guage: Conceptualizations of internal body organs across cultures and languages (pp. 349372).
Berlin: Mouton de Gruyter.
Nutton,V. (1995). Medicine in the Greek world, 80050 BC. In L. I. Conrad, M. Neve, V. Nutton,
R. Porter, & A. Wear (Eds.), The Western medical tradition: 800 BC to AD 1800 (pp. 1138).
Cambridge: Cambridge University Press.
Oster,U. (2010). Using corpus methodology for semantic and pragmatic analysis. Cognitive Lin-
guistics, 21(4), 727763.
Oxford English Dictionary Online (n.d.). Oxford: Oxford University
Padel,R. (1992). In and out of the mind: Greek images of the tragic self. Princeton, NJ: Princeton
University Press.
Partington,A. (2004). Corpora and discourse, a most congruous beast. In A. Partington, J. Morley,
& L. Haarman (Eds.), Corpora and discourse (pp. 820). Bern, Switzerland: Peter Lang.
Partington,A. (2006). Metaphors, motifs, and similes across discourse types: Corpus-Assisted
Discourse Studies (CADS) at work. In A. Stefanowitsch & S. Th. Gries (Eds.), Corpus-based
approaches to metaphor and metonymy (pp. 267304).
Partington,A. (Ed.). (2011). Modern Diachronic Corpus-assisted Discourse Studies. Corpora,

Partington,A., Morley, J., & Haarman,L. (Eds.). (2004). Corpora and discourse. Bern, Switzer-
land: Peter Lang.
Pawley,A. (1987). Encoding events in Kalam and English: Different logics for reporting experi-
ence. In R. Tomlin (Ed.), Coherence and grounding in discourse (pp. 32960). Amsterdam:
John Benjamins.
Peeters,B. (2001). Does Cognitive Linguistics live up to its name? Language and Ideology, 204,
Perkins,R. (1992). Deixis, grammar, and culture. Amsterdam: John Benjamins.
Porter,R. (1995). The eighteenth century. In L. I. Conrad, M. Neve, V. Nutton, R. Porter, & A.
Wear (Eds.), The Western medical tradition: 800 BC to AD 1800 (pp. 371475). Cambridge:
Cambridge University Press.
Porter,R. (2002). Blood and guts: A short history of medicine. New York: W. W. Norton.
Ptz,M., Niemeier, S., & Dirven, R. (Eds.). (2001). Applied cognitive linguistics I: Theory and
language acquisition. Berlin: Mouton de Gruyter.
Quinn,N. (1991). The cultural basis of metaphor. In J. Fernandez (Ed.), Beyond metaphor: The
theory of tropes in anthropology (pp. 5693). Stanford, CA: Stanford University Press.
Rauh,G. (2010). Syntactic categories: Their definition and description in linguistic theories.
Oxford: Oxford University Press.
Roob,A. (2005). Alchemy & mysticism. Cologne, Germany: Taschen.
Rosch,E., Mervis,C. B., Gray,W. D., Johnson, D. M., & Boyes-Braem, P. (1976). Basic objects in
natural categories. Cognitive Psychology, 8, 382439.
Rothstein,W. G. (1972). American physicians in the nineteenth century: From sects to science.
Baltimore, MD: Johns Hopkins University Press.
Sandra,D. (1999). What linguists can and cant tell you about the human mind: A reply to Croft.
Cognitive Linguistics, 9, 361378.
Sharifian,F., Dirven,R., Yu,N., & Niemeier,S. (2008). Culture and language: Looking for the
mind inside the body. In F. Sharifian, R. Dirven,N. Yu, & S. Niemeier (Eds.), Culture,
body, and language: Conceptualizations of internal body organs across cultures and languages
(pp. 323). Berlin: Mouton de Gruyter.
Sharifian,F., & Palmer,G. B. (Eds.). (2007). Applied cultural linguistics: Implications for second
language learning and intercultural communication. Amsterdam: John Benjamins.
Shore,B. (1996). Culture in mind: Cognition, culture, and the problem of meaning. Oxford:
Oxford University Press.
Sim,J. (2011). Metaphors of blood in American English and Hungarian: A cross-linguistic
corpus investigation. Journal of Pragmatics, 43(12), 28972910.
Sinclair,J. (1991). Corpus, concordance, collocation. Oxford: Oxford University Press.
Sinha,C., & Jensen de Lpez,K. (2000). Language, culture, and the embodiment of spatial cog-
nition. Cognitive Linguistics, 11(1/2), 1741.
Slingerland,E. (2004). Conceptions of the self in the Zhuangzi: Conceptual metaphor analysis of
comparative thought. Philosophy East and West, 54, 322342.
Starobinsky,J. (1960). Histoire du traitement de la mlancolie des origins 1900. [A history of the
treatment of melancholy, from its origins to A. D. 1900]. Basel, Switzerland: J. R. Geigy.
Steen,G. (1994). Understanding metaphor in literature. London: Longman.
Stefanowitsch,A. (2004). HAPPINESS in English and German: A metaphorical-pattern analysis.
In M. Achard & S. Kemmer (Eds.), Language, culture, and mind (pp. 134149). Stanford,
Metaphor across Time and Conceptual Space

Stefanowitsch,A. (2006a). Corpus-based approaches to metaphor and metonymy. In A.

Stefanowitsch & S. Th. Gries (Eds.), Corpus-based approaches to metaphor and metonymy
(pp. 116). Berlin: Mouton de Gruyter.
Stefanowitsch,A. (2006b). Words and their metaphors: A corpus-based approach. In A.
Stefanowitsch & S. Th. Gries (Eds.), Corpus-based approaches to metaphor and metonymy
(pp. 63105). Berlin: Mouton de Gruyter.
Stefanowitsch,A., & Gries,S. Th. (Eds.). (2006). Corpus-based approaches to metaphor and me-
tonymy. Berlin: Mouton de Gruyter.
Stubbe,H. (1662). The indian nectar, or, a discourse concerning chocolata. London: J. C. for
Andrew Crook.
Svanlund,J. (2007). Metaphor and convention. Cognitive Linguistics, 18(1), 4789.
Sweetser,E. (1990). From etymology to pragmatics: Metaphorical and cultural aspects of semantic
structure. Cambridge: Cambridge University Press.
T.,C. (1615). An advice hovv to plant tobacco in England. London: Nicholas Okes.
Taylor,J. R. (1993). Some pedagogical implications of cognitive linguistics. In R. A. Geiger & B.
Rudzka-Ostyn (Eds.), Conceptualizations and mental processing in language (pp. 201223).
Berlin: Mouton de Gruyter.
Tomasello,M. (Ed.). (1998). The new psychology of language: Cognitive and functional approach-
es to language structure, Vol. 1. Mahwah, NJ: Erlbaum.
Tomasello,M. (Ed.). (2003). The new psychology of language: Cognitive and functional approach-
es to language structure, Vol. 2. Mahwah, NJ: Erlbaum.
Traugott,E. C., & Dasher,R. B. (2002). Regularity in semantic change. Cambridge: Cambridge
University Press.
Trim,R. (2011). Metaphor and the historical evolution of conceptual mapping. London: Palgrave
Tummers,J., Heylen,K., & Geeraerts, D. (2005). Usage-based approaches in Cognitive Linguis-
tics: A technical state of the art. Corpus Linguistics and Linguistic Theory, 1(2), 225261.
Tyler,A., & Evans,V. (2001). The relation between experience, conceptual structure and mean-
ing: Non-temporal uses of tense and language teaching. In M. Ptz, S. Niemeier, & R.
Dirven (Eds.), Applied cognitive linguistics I: Theory and language acquisition (pp. 63103).
Berlin: Mouton de Gruyter.
Vesalius,A. (1998/1543). On the fabric of the human body. Book 1: The bones and cartilages
[Electronic version]. W. F. Richardson & J. B. Carman (trans.). San Francisco: Norman.
Retrieved May 26, 2013, from
Vicary,T. (1577). A profitable treatise of the anatomie of mans body. London: Henry Bamforde.
Virchow,R. (1940/1858). Cellular pathology. Ann Arbor, MI: Edwards Brothers.
Wear,A. (1995). Medicine in early modern Europe, 15001700. In L. I. Conrad, M. Neve, V.
Nutton, R. Porter, & A. Wear (Eds.), The Western medical tradition: 800 B.C. to A.D. 1800
(pp. 215361). Cambridge: Cambridge University Press.
Wilson,M. (1787). Multum in parvo being a new therapeutic-alphabet, or a pocket-dictionary, of
medicine, midwifery, & surgery (T. B. Chraghead, Trans.). Publisher unattributed.
Wiseman,R. (2007). Ancient Roman metaphors for communication. Metaphor and Symbol,
22(1), 4178.
Yu,N. (1995). Metaphorical expressions of anger and happiness in English and Chinese. Meta-
phor and Symbolic Activity, 10, 5982.
Yu,N. (2009). From body to meaning in culture. Amsterdam: John Benjamins.

Appendix A
Penn-Helsinki corpus

Text genres and word counts

Text genre Number of words Percentage of Corpus

Bible 134,275 7.5

Biography, autobiography 41,379 2.3
Biography, other 52,755 2.9
Diary, private 123,106 6.9
Drama, comedy 120,428 6.7
Educational treatise 113,032 6.3
Fiction 116,494 6.5
Handbook, other 112,419 6.3
History 108,706 6.1
Law 115,863 6.5
Letters, non-private 59,868 3.3
Letters, private 116,915 6.5
Philosophy 85,107 4.7
Proceedings, trials 105,090 8.4
Science, medicine 41,786 2.3
Science, other 79,050 4.4
Sermon 97,400 5.4
Travelogue 123,337 7.0

Totals 1,794,010 100.0

Appendix B
ARCHER corpus

Text genres and text counts

Text genre Number of texts Percentage of Corpus

Journals 100 8.1

Letters 275 22.2
Fiction, prose 100 8.1
News 100 8.1
Legal (American only) 57 4.6
Medicine (No 18th Century American) 90 7.3
Science (British only) 70 5.7
Drama (only 5 texts from 18th Century 95 7.7
Fiction, dialogue 100 8.1
Sermons 50 4.0
Court testimony 5 0.4
Essays (18th Century only) 96 7.7
Letters, Samuel Johnson 79 6.4
Prose, Samuel Johnson 21 1.7

Totals 1,238 100.1*

*Note: Percentage totals more than 100% due to rounding.

A black bile (or melancholy) 100 cognitive conceptualization

Ackerknecht 116, 120 101, 111112, 119, 126, 131, 143, (or construal) 67, 1719, 40
agitation 4243, 152, 182 146, 153, 193194, 202 dynamic vs. static 1719,
ambiguous utterance 1921, 23 Blood anger Prototype 202204
resolution 21 Scenario 5152 see also pre-cultural
ambition 140, 192 bloodletting 120, 121124, 129 Cognitive Construction
anger 31, 3637, 3956, cupping 112 Grammar 14, 16
5861, 6671, 7478, 8081, 85, blood metaphor 39, 54, 59, 77, cognitive-functionalism
9193, 9598, 136155, 157161, 85, 91, 142, 143, 145146, 147, (CF) 35, 205213, 218
169175, 184187, 191194, 202, 154155, 160161, 192, 207, 212 cognitive preference 8, 9, 200
205207 see also CM of blood Cognitive Sociolinguistics 208,
anger is a hot fluid in body heat 4243 211
a container 40, 44, 55, body is a container for cognitive style
66, 69 emotion, the 44 see pressure of coherence
anger is fire 44 Boers 74 model
anger is heat 4344, 46, 47, Boorde 100, 119, 148 cold blood 143145, 147149,
55, 6667, 68 brain 101, 112113 151, 152, 155
anger is up 155 bubbling liquid 177180, 182, cold grudge 152
see also container 184187 color 74
see also non-prototypical Burton 100, 107, 115, 152 common ground 1922,
anger Bybee 31, 6365 204205
anger Prototype Scenario 47 competence 45, 215
48, 5153 C see also performance
angry person is a C.T. 121 compiled corpus
pressurized container, Cameron 18, 19, 58 see corpus
the 70, 74 category complaint 151
A Representative Corpus of classical 5 complementary
Historical English Registers natural 5 distribution 192, 207
(ARCHER) 9091, 150159, Casasanto 29 conceptual metaphor (CM) 3,
231 cell pathology 125 3940, 196197
choler 100101, 111113, 119, dimension 4950, 9192, 144,
B 131, 139 192193, 212
background event 4849 choleric elaboration 4346
see also presupposed event see Four Humors model entailment 63, 142, 145, 150,
balance (of the humors) 107, Chomsky 45, 199 152, 157, 192, 193
111113, 115116, 123124 Cienki 48, 59 Conceptual Metaphor Theory
Barcelona 58, 60 circulation (of the blood) 118, (CMT) 7, 3940, 196197
Barlow 1517 120, 136 conceptual unity (of do-
base 67, 40, 4849, 74 Clark 1921 main) 48, 203, 215
see also profile; unprofiled CM of anger 4244, 46, 4954, constraint (on mapping) 5657
base 8991, 191192 container 44, 45, 46, 55, 72,
Bertuol 6465 CM of blood 39 142, 145, 147, 149, 156, 159, 168,
Biber 29, 9798, 209 CM of spleen 39 169, 181
Metaphor across Time and Conceptual Space

context 2628, 93, 168, 208 cultural periphery 199, 200 effervescence 158
context-induced metaphor 5758 cultural unit 9, 14, 48 elaboration
control 5253, 55, 148152, culture 7, 7475, 80 see conceptual metaphor
192193 local knowledge 7, 9 see lexical elaboration
see also scale see also cultural knowledge Emanatian 5960
controlled response over embodied core 199, 200, 203,
time 4950, 151 D 204, 213
cooking semantic frame 44 DAndrade 10 embodiment 9, 15, 5660,
45, 149, 150, 156160, 194, Dariot 105 6667, 69, 77, 146, 168, 193,
205, 207 de Mediolano 100 198202, 204, 206, 213
cool blood 148, 154 deictic orientation 1114 embodied experience 8, 9, 11,
coordination device 1921, 23 orientation type 12 12, 14, 15, 4043, 56, 60, 66,
core meaning face-to-face 12 67, 70, 78, 79, 192, 193, 195,
see linguistic metaphor single-file 12 198205, 207, 212, 217
core vs. periphery 5 Deignan 23, 5354 embodied realism 40, 56, 57
see also embodied core; diachronic embodiment principle 56
cultural periphery cultural model 39, 47, 6365, emergent 17, 198
corpus 23, 2935, 59, 67, 68, 72, 77 emotion
78, 80, 86, 90, 91, 94, 95, 97, language change 16, 75 see domain matrix
98, 109, 117, 119121, 123, 124, metaphor 6367, 6877, emotions are fluids in a
126, 133, 134, 137139, 141, 217218 container 73
150, 159, 167, 178181, 187, 195, research 25, 36, 65, 67, 8081, empirical 1617, 23, 27, 118
208210, 213, 217, 229, 231 207208 encyclopedic knowledge 7, 192
compiled 29, 30, 34, 89, 92, see also time energy 156
94, 138, 153, 166, 194, 208 Diller 75, 76 Enfield 1314, 2425, 48, 196
short-term language forms dimension entailment
vs. long-term language see conceptual metaphor see conceptual metaphor
patterns 30, 33 see scale entity 72
representative 90, 195 Dirven 10, 208, 216, 217 envy 151
research methods 2829, discourse analysis 34, 98, 137, Estes 126
3033 192 eucrasia 116
size 209210 dislike 147 see also balance
text selection 30, 32 dissection 112, 117118 event typicality 13, 196
see also text corpora domain 67, 28, 31, 9596, excitation 180, 182184
Corpus-assisted Discourse 202204 experience, human 5, 5658,
Studies (CADS) 3436, unity of, 202 207, 218
9799, 210 domain matrix 1011, 91, 140, experiencer 11
Corpus of Historical American 174, 177, 206, 212, 213 experiential scene 12, 14, 4751,
English (COHA) 209210 of emotion 11, 37, 78, 178, 53, 5657, 197, 206, 212
courage 168 186, 192195, 202, 204, 206 explicit target
Croft 7, 1011, 1920, 2728, see also conceptual unity (of see target domain
202203 domain) explosion 45, 170
Cuff 100, 105106, 110 Domaradzki 15, 77 expression (of emotion)
cultural knowledge 79, Draper 87 private 75, 76, 77, 78, 147, 151,
198199, 203, 210, 213 dynamic construal 1719, 183, 192, 207
culturally specific embodiment 198204, 205206 public 75, 76, 77, 192, 207
(CSE) 56 dynamism 198, 199, 202
culturally tainted dyscrasia 116, 125126, 195 F
embodiment 56 see also balance Father Time 198
cultural model 710, 1115, fear 145, 155
5660, 86, 206207, 217 E Fesmire 2627
cognitive characteristic 10, ebullition 170 Fillmore 67, 48
11, 12, 14 ecology principle 27 Flexner Report 127129
see also mental model Eerden 26 flowing 126, 182

fluid CM 4446 heat 41, 44, 45, 47, 66, 69, 143, introspection 2729, 31, 32, 80
fondness 154, 192 144, 147, 149, 150, 170171, 185,
foot 63 193, 205 J
force 73, 149 see also scale Jackson 123
form/meaning pair 19 Heine 1112, 14 jealousy 151
Four Humors model 100101, Heylen 29 Johnson, M. 2526, 40, 181,
110115 Hiatt 128 182, 184
humors 100101, 111112 Historical Four Humors joint salience, principle of 21
organs 100101, 112 Texts 8788, 100101, joy 142, 192
qualities 100101, 110 219220 justice 141, 143
temperaments 114115 hot/cold 100101, 110, 148, 150
choleric 100101, 115 Huarte 101, 105, 112 K
melancholic 100101, 115 human body 7273 Keesing 8
phlegmatic 100101, 115 human mind 36 key 3, 11, 2021
sanguine 100101, 114115 human physiology 8990, 116, keyword search procedure 93,
frame 67, 40 117 167
frenzy 182, 185 Hunston 30 kill 141
frequency hunting 141 Klibansky 119
analysis 68, 98, 160 Koivisto-Alanko 7273, 9394
of use 3132, 5960, 9698, I Kvecses 89, 33, 4244, 4648,
196 idealized cognitive model 4951, 6971, 168175, 187,
statistic 7980 (ICM) 6, 40 200202
illness 147, 174175 Kristiansen 208
G image/attribute metaphor
Galen 117118, 120 see resemblance metaphor L
Geeraerts 28, 3031, 6566, image schema Lakoff 3940, 4257, 157,
7577 see container 168169, 173
Generative Grammar 45 imagination 151152, 192 Langacker 6, 1011, 1617, 40,
genre 29, 75, 76 implicit target 198200, 203, 206
Gentner 10 see target domain language 36
Gevaert 6869 inflammation 123, 124 language change 74, 75
Ghesquire 22, 67 insanity 153, 174 language universal
Gibbs 28 inspiration 182 see universal
Goldberg 12, 1314, 16, 48 instrument/tool/weapon 72 lesser world frame 105, 110111
Grace 13 intense response over lexical elaboration 44
Grady 5758, 143 time 5052, 145, 195 lexical field analysis 79
great world frame 105106, intensity 145 lexicalized target
110111 of activity is heat 143 see target domain
Green, G. M. 1921, 205 of emotion is heat 143 linguacentrism 2425, 32
Green, M. 116, 123 of emotion is intensity linguistic expression 12, 1617
grief 146 of motion 143, 145, 156, linguistic metaphor 23, 5758,
Gries 29, 31, 211 159, 186 9596
grudge 50, 152 of energy is intensity of core meaning 23
motion 156, 158 non-core use 23
H of motion 143, 145 Linguistic Relativity Hypothesis
happiness 108, 153, 177178, interference with accurate (LRH) 24
180187 perception 42 linguocentrism
happiness is a liquid 177 internal pressure 42, 43 See linguacentrism
happiness is a liquid in a Internet Library of Early Jour- liquid in a container 181,
container 181 nals (ILEJ) 166 184
Harvey 120122 interpretation Littlemore 216
hate 192 see construal; perspective liver 100101, 112, 113, 118, 119
heart 7071, 100101, 110111, intersubjectification 22 local knowledge
120, 146, 155, 184 intersubjective 6, 14, 22, 4142 see culture
Metaphor across Time and Conceptual Space

longitudinal research N physiological effects of an

see diachronic research natural heat 106, 110111, 114, emotion stand for the
long-term language patterns 139, 148 emotion, the 42, 43
see corpus natural language physiology, animal 117118
love 73, 74, 192 data 91, 166 physiology, human
love is a valuable use 29, 61 dissection 112, 117118
commodity 72 New York Times 129 scientific advances 86, 89, 90,
Lucy 2425, 210 Niemeier 10, 13, 213, 216 94, 103, 108, 109, 116131,
Ludmerer 127 Ninteenth Century in Print 133, 136, 160, 161, 194
lungs 100, 113, 119120 (NCP) 166 Porter 118, 122
Lutz 69 non-autonomous knowledge 7, pragmatics 4, 5
41, 77, 204, 206, 216 pre-cultural 6, 40, 41, 79, 198
M non-core use see also cognitive conceptual-
Maalej 5657, 60, 145 see linguistic metaphor ization (or construal)
MacArthur 5859 non-linguistic data 2426, pressure 42, 43, 4546, 130, 131,
macrocosm/microcosm 36, 72, 8690, 103104, 208, 168171, 173, 180, 181182, 193
model 105107 210211 pressurized container 70,
macrocosmos 105, 106 non-linguistic knowledge 8, 18 71, 74
microcosmos 105, 106 non-objectivist 30, 33, 212 see also internal pressure
macro-study 37, 77, 83 non-prototypical anger 4951, pressure of coherence mod-
madness is an illness 174 5253, 145 el 89, 17, 79
Matsuki 5455, 199 normal belief 21, 205 see also override
medical sects 128129 Normalized Frequency Rate presupposed event 48, 49
Eclectic 129 (NFR) 97, 98, 160 see also background event
Thomsonian 129 Nutton 87 primary CM 70, 143
melancholic profile 4849, 141142
see Four Humors model O progress is forward mo-
melancholy adust 153, 193 obstacle 72 tion 35
mental model 10 ontological metaphor 72 protection 75
mental space 202, 203 orientational metaphor 35 prototype 78, 154155, 191192,
metaphoreme 19 Oster 211 212
metaphorical pattern 92, 95 override 9, 200201, 204205 prototypicality 31, 80, 137138,
metaphorical expression 9295, see also pressure of coherence 161, 197, 207
143, 204, 210 model purgation 120, 123
metaphoric property search Oxford English Dictionary Ptz 216
technique 178, 187 Online (OED-O) 45, 143, 169
micro-study 77 Q
mind is a container, the 72 P qualitative analysis 98, 211
Mischler 59 Padel 112 Quinn 15
mixed-methods research 208, Partington 28, 3436, 211
210, 211 Peeters 29 R
Moder 166 Penn-Helsinki Parsed Corpus Rauh 12
moist (or wet)/dry 100101, 110, of Early Modern English reason 72, 73, 144, 146, 148152,
148149 (PPCEME) 90, 141150, 230 154, 194, 196, 202, 212
Morgagni 122, 136 performance 45, 215 see also scale
Morley 34 see also competence reference corpus 35
morphology 5, 206 Perkins 10, 165 register 29, 90, 195
movement 156 personification 73, 96, 138, 141, relational metaphor
moving forward is neces- 144145, 198 see resemblance metaphor
sary 35 perspective (on a scene) 14, 41 relativism 11, 15
Mller 19, 26 phlegm 100101, 111, 119 Reppen 29
multidisciplinary research 59, phlegmatic research
60, 208 see Four Humors model future 205208
multimodal metaphor 26, 211 phonology 5, 206 methodology 208212

synchronic-historical 64, society is a family CM 201 Tomasello 3, 4

65, 209 source domain 31, 43, 44, 5658, token 68, 72
see also diachronic 72, 73, 92, 93, 9596 Traugott 22
resemblance metaphor 5758 see also domain Trim 32, 33, 74, 75, 77, 80, 196,
image/attribute 58 speech community 6, 7, 8, 9, 213
relational 58 1214, 1924, 3234, 205207 truth conditions 5
retribution 47, 49, 51, 53, spleen 4546, 165175, 177, 183, Tummers 17
149151, 154, 183 186 Tyler 217
revenge 141, 192 Spleen anger Prototype Sce-
Roob 119 nario 51, 53 U
Rosch 5 spleen metaphor 36, 46, 49, 54, unified model 104, 107108,
Rothstein 129 77, 85, 91, 146, 147, 192193, 109, 118120, 120122, 123125,
199, 207 126127, 129130
S see also CM of spleen universal 69, 14, 22, 28, 40,
sadness 96, 124, 127, 143, 146, 147, Steen 54 41, 70, 74, 77, 79, 198, 199,
150, 152, 172, 180, 184185, 192 Stefanowitsch 29, 92, 93, 95, 166, 201, 203
salience (or salient) 19, 22 177, 211 cross-cultural universal 7
Sandra 29 stratification 198 universalism 11, 15
sanguine Stubbe 121 universalist vs. relativist 77
see Four Humors model support 63 universe is mathematics 64
scale 45, 144, 146, 149, 154, 155, Sweetser 60, 63, 165, 208 unprofiled base 48
171, 173, 180, 193 symbolic unit 4, 5, 19 see also base
schema 10, 16, 17, 55, 73, 142, 168, synchronic-historical Usage-based theory of lan-
185, 202, 216 see research guage 15, 16
scientific advances syntax 5, 1114, 41, 206 usage event 16, 216
see physiology, human systematic metaphor 58 utterance 5, 18, 20
selection (grammatical catego-
ries) 202 T V
self-care focus 104, 108 target domain 11, 31, 40, 5657, variation (in conceptualiza-
see also Four Humors model 72, 73, 92, 93, 9596 tion) 158, 200
semantic field 23, 70, 74, 80 implied 92, 93, 9596 vengeance 138, 145
semantic frame 14 lexicalized (or explicit) 92 verb serialization 13, 48
semantics 4, 5 see also domain Vesalius 117118, 122
semantic shift 196, 197 temperature 66, 113114 Vicary 113, 119, 138, 139
semiotic triangle 19, 196, 208, temperature 144, 146, 148155, Virchow 116, 122, 125126, 136,
216 157, 171172, 180, 181, 195, 212 160, 195
sexual attraction 192 see also scale
Sharifian 10, 216 text corpora 23, 2729, 90, 92, W
shield 74, 75 93, 211 war 74
Shore 10 text mode 29 Wear 117, 118
short-term language forms text selection Wilson 123
see corpus see corpus wit 73
Sim 5960, 148 text type 75, 76 wrath 51, 142, 145, 155156
Sinclair 28 theory of language 4, 1517,
Sinha 10, 15, 213 3031 Y
skin redness 42, 43, 55, 130, 131, time 3, 23, 6365 Yu 10, 5556, 65, 68, 199, 213
152, 168, 171, 173 time 4951, 198