analysis of conceptual metaphors in the

University of Niš

FACULTY OF PHILOSOPHY

MASTER THESIS

Vladimir Figar

University of Niš

FACULTY OF PHILOSOPHY

English Department MA Studies

Analysis of Conceptual Metaphors in the Political Discourse of Daily Newspapers: Structure, Function, and Emotional Appeal

Mentor Student Mihailo Antović, Ph. D. Vladimir Figar, 39

Niš, 2013

Acknowledgment

I am grateful to my mentor, Mihailo Antović, Ph.D. for his support, critical comments, and suggestions. I would also like to thank Todd Oakley, Ph.D. for his critical remarks pertaining to the first part of the present research. Additionally, I would like to express my gratitude to all of my colleagues from the English Department who participated in the second part of the research and made this study possible. Finally, I thank my family and friends for their support and patience.

i

Table of Contents

Table of Contents ________________________________________________________________________i 1. Introduction__________________________________________________________________________ 9 2. Theoretical Framework _______________________________________________________________ 11

2.1. Conceptual Metaphor Theory (CMT) _________________________________________________ 11 2.1.1. Structural, Orientational, and Ontological Metaphors __________________________________ 12 2.1.2. Systematicity and Partial Nature of Metaphorical Mappings _____________________________ 12 2.1.3. Metaphorical Entailments _______________________________________________________ 13 2.1.4. Metaphor Systems and Metaphorical Coherence _____________________________________ 13 2.1.5. Image Schemas and the Reality behind Them _______________________________________ 15 2.1.6. The Invariance Principle ________________________________________________________ 16 2.1.7. Frames and ICMs _____________________________________________________________ 17 2.1.8. Conventional and Novel Metaphors________________________________________________ 18 2.1.9. Critical Metaphor Analysis (CMA) _________________________________________________ 19 2.1.10. Metaphor and Politics _________________________________________________________ 20 2.1.11. Section Summary ____________________________________________________________ 23

2.2. Conceptual Blending Theory (CBT) _________________________________________________ 24 2.2.1. Mental Spaces and Backstage Cognition ___________________________________________ 24 2.2.2. Conceptual Blending ___________________________________________________________ 26 2.2.3. The Generic Space Problem _____________________________________________________ 27 2.2.3. Vital Relations ________________________________________________________________ 29 2.2.4. Compression _________________________________________________________________ 31 2.2.5. Optimality Principles ___________________________________________________________ 33 2.2.6. Basic Types of Conceptual Integration Networks _____________________________________ 35

2.2.6.3. Single-scope Networks______________________________________________________ 35 2.2.6.4. Double-scope Networks _____________________________________________________ 37

2.2.7. CBT and Metaphor ____________________________________________________________ 39 2.2.8. CMT vs. CBT: Towards a Unitary Framework ________________________________________ 41 2.2.9. Taking a Slightly Different Turn: The Coded Meaning Model ____________________________ 43 2.2.10. Section Summary ____________________________________________________________ 45

2.3. Affect and Emotions ______________________________________________________________ 46 2.3.1. Emotion Concepts _____________________________________________________________ 46 2.3.2. Affect _______________________________________________________________________ 47

2.3.2.1. The Circumplex Model of Affect _______________________________________________ 47 2.3.2.2. The PANAS Model _________________________________________________________ 49 2.3.2.3. The Core Affect Model ______________________________________________________ 51

2.3.2.3.1. Core Affect and Cognition________________________________________________ 52 2.3.3. The Conceptual Act Theory of Emotion_____________________________________________ 53 2.3.4. Conceptual Act Theory of Emotion and Conceptual Blending Theory: Points of Intersection____ 55

ii

2.3.5. Affect, Emotions, Media, and Politics ______________________________________________ 57 2.3.6. Section Summary _____________________________________________________________ 58

3. Previous Research ___________________________________________________________________ 59 3.1. CMT Framework _________________________________________________________________ 59 3.2. CBT Framework__________________________________________________________________ 60 3.3. Political Psychology ______________________________________________________________ 63 3.4. Section Summary ________________________________________________________________ 63

4. Present Research ____________________________________________________________________ 65 4.1. Present Research: Part 1 __________________________________________________________ 66

4.1.1. Corpus and Methodology________________________________________________________ 66 4.1.2. Corpus Analysis and Discussion __________________________________________________ 67

4.1.2.1.1. Structure of CONFLICT Metaphors __________________________________________ 68 4.1.2.1.2. Systematicity of Emergent Structure in CONFLICT Metaphors _____________________ 85 4.1.2.1.3. Function and Emotional Appeal of CONFLICT Metaphors_________________________ 88 4.1.2.1.4. Metaphor Systems and Metaphor Interaction_________________________________ 89

4.1.2.2. SPORT Metaphors __________________________________________________________ 91 4.1.2.2.1. Structure of SPORT Metaphors_____________________________________________ 91 4.1.2.2.2. Systematicity of Emergent Structure in SPORT Metaphors ______________________ 105 4.1.2.2.3. Function and Emotional Appeal of SPORT Metaphors __________________________ 107 4.1.2.2.4. Metaphor Systems and Metaphor Interaction________________________________ 108

4.1.2.3. Network Optimization in CONFLICT and SPORT Metaphors___________________________ 110 4.2. Present Research: Part 2 _________________________________________________________ 114

4.2.1. Questionnaire 1: Results and Discussion __________________________________________ 116 4.2.1.1. Scale Reliability __________________________________________________________ 116 4.2.1.2. Assumptions Required for the Application of Parametric Tests _____________________ 117 4.2.1.3. Means and One-Sample T-Test ______________________________________________ 118

4.2.2. Questionnaire 2: Results and Discussion __________________________________________ 119 4.2.2.1. Scale Reliability __________________________________________________________ 119 4.2.2.2. Means and One-Sample T-Test ______________________________________________ 120

4.2.3. Questionnaires 1 and 2: Comparison of Findings ____________________________________ 122 4.2.4. Questionnaire 3: Results and Discussion __________________________________________ 124

4.2.4.1. Scale Reliability __________________________________________________________ 124 4.2.4.2. Means and One-Sample T-Test ______________________________________________ 124

4.2.5. Questionnaire 4: Results and Discussion __________________________________________ 125 4.2.5.1. Scale Reliability __________________________________________________________ 125 4.2.5.2. Means and One-Sample T-Test ______________________________________________ 126

4.2.6. Questionnaires 3 and 4: Comparison of Findings ____________________________________ 128 5. General Discussion _________________________________________________________________ 131

5.1. Research Questions 1 through 6___________________________________________________ 131 5.2. Research Questions 7 and 8 ______________________________________________________ 134 5.3. Parts 1 and 2 of the Present Research – Points of Intersection__________________________ 135 5.4. Additional Remarks _____________________________________________________________ 139 5.5. Section Summary _______________________________________________________________ 141

6. Conclusions and Suggestions for Further Research ______________________________________ 142 References __________________________________________________________________________ 144

iii

Appendix B – Questionnaire 2 __________________________________________________________ 152 Appendix C – Questionnaire 3 __________________________________________________________ 154 Appendix D - Questionnaire 4 ___________________________________________________________ 156

iv

“Metaphor is a solar eclipse. It hides the object of study and at the same time reveals some of its most salient and interesting characteristics when viewed through the right telescope.” (Paivio & Walsh, 1993: 307)

v

Analysis of Conceptual Metaphors in the Political Discourse of Daily Newspapers: Structure, Function, and Emotional Appeal

Abstract: The present paper explores the structure, function, and the potential of conceptual metaphors in the political discourse of daily newspapers to provoke an emotional reaction with the electorate, i.e. readers. The two groups of metaphors that are investigated include CONFLICT and SPORT metaphors. Consequently, the corpus consists of metaphorical expressions corresponding to the two previously mentioned conceptual keys, extracted from the on-line editions of The New York Times during December 2011. The paper resides on an integrated theoretical framework that encompasses the basic tenets of the Conceptual Metaphor Theory (CMT), the fully developed model of the Conceptual Blending Theory (CBT), and the Conceptual Act Theory of Emotion. Additionally, basic tenets of the Circumplex Model of Affect and the Positive and Negative Affect Schedule (PANAS) are also utilized. The present research is conducted in two stages: 1) metaphorical expressions from the corpus are analyzed in terms of their structure and function through the concomitant use of the CMT and CBT models; 2) building on the Conceptual Act Theory of Emotion, Russell’s circumplex, and the PANAS model, the second stage of the research uses appropriate questionnaires to test whether the metaphorical expressions extracted from the corpus can provoke an actual tangible emotional reaction with the experimental subjects. The reaction is measured in terms of affect, and in terms of specific emotion concepts. Such research design is meant to reveal both the cognitive mechanisms that underlie the construction of metaphorical conceptual integration networks, and the mechanisms that facilitate the development of an emotional response. Furthermore, the results obtained from the two parts of the present research are compared and investigated for potential points of intersection. Additionally, the results of the present study are also compared against the results obtained from the previous research in the field, predominantly conducted within the CMT framework. Finally, the paper also presents additional evidence in favor of the joint use of the CMT and the CBT frameworks. Based on the analyses, it can be concluded that all metaphorical conceptual integration networks from the present corpus appeared as single-scope networks. Additionally, a high degree of systematicity of emergent structures was also recorded, which was accounted for by the highly entrenched nature of the two conceptual keys, and links were established with the systematic nature of compression. The two main cognitive mechanisms that facilitate the creation of an emotional reaction are backward

projections and coupled elaboration, which in turn enable metaphors to function as powerful persuasive tools. The second part of the research revealed that metaphorical expressions from the corpus can provoke a certain degree of an emotional response with the readers, i.e. experimental subjects. In addition,

vi

the recorded reactions also showed a certain degree of consistency. In conclusion, metaphor poses as a powerful rhetorical tool that can influence individuals’ reasoning and behavior by influencing their emotional and affective states. Future research should attempt to further reveal the nature of interaction between the processes of meaning and emotion construction. Key words: conceptual metaphor, conceptual blending, compression, emergent structure, network optimization, human scale, political discourse of daily newspapers, emotional response, affect

vii

Analiza konceptualnih metafora u političkom diskursu dnevnih novina: struktura, funkcija i emocionalni efekat

Apstrakt: Ovo istraživanje bavi se strukturom, funkcijom i sposobnošću konceptualnih metafora u političkom diskursu dnevnih novina da izazovu emocionalnu reakciju kod publike. Dve grupe metafora koje će u radu biti analizirane uključuju metafore KONFLIKTA i SPORTSKE metafore. Shodno tome, korpus se sastoji od metaforičkih izraza koji odgovaraju prethodno spomenutim konceptualnim ključevima (conceptual keys) koji su dobijeni iz onlajn izdanja dnevnih novina New York Times tokom decembra meseca 2011. godine. Rad je zasnovan na teorijskom modelu koji obuhvata osnovne smernice teorije

konceptualne metafore Lejkofa i Džonsona (Conceptual Metaphor Theory), teoriju konceptualnog stapanja Fokonijea i Tarnera (Conceptual Blending Theory) i teoriju emocija kao konceptualnog čina (Conceptual

Act Theory of Emotion). Pored toga, rad se delimično oslanja i na Raselov cirkumpleksni model afekta (Circumplex Model of Affect), kao i na PANAS model (Positive and Negative Affect Schedule). Istraživanje je obavljeno u dve etape: 1) analiza strukture i funkcije metaforičkih izraza iz korpusa kroz teorijski model teorije konceptualne metafore i teorije konceptualnog stapanja; 2) druga etapa istraživanja oslanja se na odgovarajuće upitnike posredstvom kojih se ispituje da li metaforički izrazi iz korpusa mogu da izazovu ’opipljivu’ emocionalnu reakciju kod učesnika u istraživanju. Ova reakcija meri se u odnosu na afekat i u odnosu na konkretne koncepte koji opisuju emocije (emotion concepts). Ovakav nacrt istraživanja ima za cilj da otkrije kognitivne mehanizme koji omogućavaju formiranje metaforičkih mreža konceptualne integracije, kao i mehanizme koji omogućavaju razvoj emocionalne reakcije. Takođe, rezultati dobijeni u dve etape ovog istraživanja biće upoređeni najpre međusobno, a potom sa rezultatima prethodnih istraživanja iz ove oblasti. Pored toga, ovo istraživanje će pokušati da predstavi dodatne argumente u prilog zajedničkoj upotrebi teorije konceptualne metafore i teorije konceptualnog stapanja. Na osnovu rezultata analize korpusa može se zaključiti da su se sve metaforičke mreže konceptualne integracije pojavljivale kao mreže jednostrukog opsega. Pored toga, ustanovljen je i visok stepen sistematičnosti u formama emergentnih strucktura, što je objašnjeno visokim nivoom konvencionalnosti konceptualnih ključeva. Takođe, uspostavljena je i veza između sistematičnosti u formi emergentnih struktura i sistematske prirode kompresije. Dva glavna kognitivna mehanizma koji omogućavaju izazivanje emocionalne reakcije su projekcije unazad (backward projections) i uparena elaboracija (coupled

elaboration). Ovi mehanizmi takođe omogućavaju metaforama da obavljaju svoju retoričku funkciju. Drugi deo istraživanja otkrio je da metaforički izrazi iz korpusa zaista mogu da izazovu određeni stepen emocionalne reakcije kod ispitanika. Pored toga, zabeležene reakcije takođe su pokazale određeni stepen doslednosti. Na osnovu rezultata dobijenih u ovom istraživanju može se zaključiti da metafora predstavlja moćno retoričko oružje koje može uticati na rezonovanje i ponašanje pojedinaca tako što utiče na njihovo

8

afektivno i emocionalno stanje. Dalja istraživanja trebala bi da otkriju više detalja o prirodi interakcije između procesa konstrukcije značenja i konstrukcije emocionalnog iskustva. Ključne reči: konceptualna metafora, konceptualno stapanje, kompresija, emergentna struktura, optimizacija mreža, politički diskurs dnevnih novina, emocionalna reakcija, afekat

Introduction 9

1. Introduction

With the onset of cognitive linguistics, metaphor was transformed from a once anomalous phenomenon, to an indispensable asset in the human cognitive arsenal. Additionally, cognitive linguists claim language to be a function of our cognitive systems, rather than an inborn capacity waiting to unravel, rendering it largely dependent on the bodily experience of the world around us (Lakoff & Johnson, 1980a; 2003[1980b]; Lakoff, 1991; 2006[1993]; Saeed, 2003; Evans & Green, 2006; Gibbs & Colston, 2006[1995]; Mandler, 1992; 2012). In other words, “cognitive semanticists take the view that we have no access to a reality independent of human categorization” (Saeed, 2003: 344), which means that “our conceptual system ... plays a central role in defining our realities” (Lakoff & Johnson, 2003[1980b]: 3). Bearing in mind that “our ordinary conceptual system ... is fundamentally metaphorical in nature” (Lakoff & Johnson, 2003[1980b]: 3), it is easy to understand how important conceptual metaphor is in everyday communication.

The main aim of the present research will be to explore the structure and function of conceptual metaphors in the political discourse of daily newspapers. Additionally, the paper will also investigate whether these metaphors can in fact provoke an actual, tangible emotional response with the readers. In that sense, the present research will be divided into two integral parts. The results obtained from these two stages of the research will be then compared and investigated for potential points of intersection. The specifics pertaining to the theoretical framework, corpus construction, and the methodologies used in the analyses are described in more detail in sections 2.1, 2.2, 2.3, 4.1, and 4.2. The paragraphs below provide a general overview of the structure of the paper, as well as the information pertaining to the content of its main sections.

The paper will begin with the introduction of the theoretical framework (section 2) on which the present research will be grounded. Section 2.1 will introduce the basic notions pertaining to the Conceptual Metaphor Theory (CMT), and it will also address some attention to the notions of image schemas (subsection 2.1.5), and frames and ICMs (subsection 2.1.7), which will be essential for further analyses. Furthermore, subsection 2.1.9 will introduce the basic elements of the Critical Metaphor Analysis as presented in Charteris-Black (2004), while subsection 2.1.10 will be dealing more explicitly with the role of conceptual metaphors in political discourse.

Section 2.2 will in turn introduce the theoretical framework of the Conceptual Blending Theory (CBT) as presented in Fauconnier and Turner (2002; 2006[1998]). Namely, the section will open with the discussion of the basic tenets of the Mental Space Theory (Fauconnier, 1994; 1997; 2007), after which it will proceed to its logical sequel – the CBT model. The discussion of the CBT model will include some important notions that will later be utilized in the analyses of the examples from the corpus. Some of these theoretical concepts encompass the notions of vital relations, compressions, systematic nature of compression, optimality principles, and single- and double-scope networks. Subsection 2.2.8 will pay special attention to the relationship between

Introduction 10

the CMT and the CBT model, and it will foreground the arguments in favor of the joint use of the two theoretical frameworks. In summary, sections 2.1 and 2.2 will comprise an integrated theoretical framework that will be used in the first part of the present research. In addition to the now canonical blending model of Fauconnier and Turner (2002), subsection 2.2.9 will also present an outline of a more recent blending framework, the Coded Meaning Model introduced by Coulson and Oakley (2005).

Section 2.3 will be dealing with the theoretical framework that will be used in the second part of the present research. Namely, this section will introduce the basic elements of the Circumplex Model of Affect (Russell, 1980), the PANAS model (Watson, Clark, & Tellegen, 1988), and the Conceptual Act Theory of Emotion (Barrett, 2006; Lindquist & Barrett, 2008). Additionally, the constructs of core affect (Russell, 2003) and situated conceptualizations (Barsalou, 2009) will also be briefly discussed, owing to their intrinsic link to the Conceptual Act Theory.

Following the introduction of the theoretical framework, section 3 will present some relevant results from the previous research, which will be followed by section 4 that will introduce the present research and the main research questions that this paper will attempt to answer. The present research will consist of two parts. The first part, described in section 4.1, will be dealing with corpus analysis using the combined theoretical model of the CMT and the CBT frameworks. Such an approach is expected to provide a thorough insight into both the cognitive mechanisms that underlie the construction of metaphorical conceptual integration networks, and the social functions that metaphors perform in the political discourse of daily newspapers and mechanisms that facilitate those functions. Examples from the corpus will be organized around the two main conceptual keys – CONFLICT and SPORT.

Section 4.2 will introduce the second part of the present research which was designed to determine whether conceptual metaphors from the political discourse of daily newspapers can provoke an actual tangible emotional reaction with the experimental subjects. For that purpose two types of questionnaires were designed. Type-one questionnaires were developed to measure the affective response in terms of valence and arousal, while type-two questionnaires were developed to measure the emotional reaction in terms of more specific emotion concepts presented in the form of a checklist with 4 positive and 4 negative emotion concepts.

In addition to the initial discussions of the obtained results, performed in sections 4.1 and 4.2, section 5 will provide a comprehensive overview of the results obtained in both parts of the present research. Namely, this section begins with an elaborate discussion of the main research questions (subsections 5.1 and 5.2), followed by an in-depth comparison of findings from the two parts of the present research (subsection 5.3). Subsection 5.4 presents additional arguments in favor of the concomitant use of the CMT and the CBT models, and once again highlights the points of intersection between the CBT model and the Conceptual Act Theory, which license the construction of the fully developed human scale experience.

Finally, the paper closes with section 6 which provides a short overview of the most important findings from the current research, along with their implications for future research.

Theoretical Framework 11

2. Theoretical Framework The current section will outline the structure of the theoretical framework that will be used in the paper.

Namely, this will encompass the CMT model, the CBT model, the Conceptual Act Theory of Emotion, Russell’s circumplex, and the PANAS model. Some additional theoretical constructs, such as core affect and situated

conceptualizations, essential for the understanding of the Conceptual Act Theory of Emotion will also be discussed.

2.1. Conceptual Metaphor Theory (CMT)

Instead of viewing metaphor as a purely linguistic phenomenon, cognitive scientists understand it as “a pattern of conceptual association” (Grady, 2007: 188) that can also be expressed in a nonverbal manner. In other words, “the locus of metaphor is not in language ... but in the way we conceptualize one mental domain in terms of another” (Lakoff, 2006[1993]: 185). In line with such an interpretation, Conceptual Metaphor Theory

(CMT) proposes that “the essence of metaphor is understanding and experiencing one kind of thing in terms of another” (Lakoff & Johnson, 2003[1980b]: 5). With regard to its structure, conceptual metaphor is characterized by unidirectional systematic partial mappings from the source to the target domain (Lakoff & Johnson, 2003[1980b]; Lakoff, 2006[1993]; Saeed, 2003; Evans & Green, 2006; Kövecses, 2006; 2010). In CMT, the term domain refers to “a body of knowledge that organizes related concepts” (Evans & Green, 2006: 190), where “the source is a more physical, and the target a more abstract kind of domain” (Kövecses, 2006: 117). This implies that metaphor facilitates the understanding of target domains that are normally more vague and abstract, via source domains that are more tangible and concrete. Additionally, the unidirectional nature of metaphorical cross-domain mappings suggests that metaphors “provoke the listener to transfer features from the source to the target” (Saeed, 2003: 350), or in other words asymmetrically.

While such a cognitive mechanism is of tremendous value when it comes, for example, to learning strategies, it also has one major setback. Namely, “metaphor highlights certain features while suppressing others” (Lakoff & Johnson, 2003[1980b]: 141), which suggests that some aspects of a concept will necessarily remain hidden. In turn, “various metaphorical structurings of a concept serve different purposes by highlighting different aspects of the concept” (ibid.: 96). Although this can be harmless in some cases, it can also be used as a strategy that enables a ‘silent’ introduction of entire belief and value systems. Whatever the case, perhaps the greatest appeal to metaphor is situated in the fact that it goes on largely undetected.

Conceptual Metaphor Theory 12

2.1.1. Structural, Orientational, and Ontological Metaphors

Lakoff and Johnson (1980a; 2003[1980b]) provided a basic classification of conceptual metaphors that includes structural, orientational, and ontological metaphors.

Structural metaphors represent “cases where one concept is metaphorically structured in terms of another” (Lakoff & Johnson, 2003[1980b]: 14), and “the source domain provides a relatively rich knowledge structure for the target concept” (Kövecses, 2010: 37). For example, the metaphor ARGUMENT IS WAR provides one instance of a structural metaphor, where the concept of argument is structured relative to the concept of war. The cross-domain mappings involved in this metaphor do not only account for the meaning of its individual linguistic realizations (i.e. metaphorical expressions), but also for the understanding of the concept of an argument and its essence.

Orientational metaphors “give a concept a spatial orientation” (Lakoff and Johnson, 2003[1980b]: 14), which has a strong experiential basis grounded in both physical and cultural experience. In other words, these metaphors “structure concepts linearly, orienting them with respect to nonmetaphorical linear orientations” (Lakoff & Johnson, 1980a: 195). Lakoff and Johnson (2003[1980b]: 14-21) list the following pairs of orientational metaphors: HAPPY IS UP – SAD IS DOWN; CONSCIOUS IS UP – UNCONSCIOUS IS DOWN; MORE IS UP –

LESS IS DOWN; GOOD IS UP – BAD IS DOWN; etc. According to these examples, it is obvious that upward orientation is coupled with a positive evaluation, whereas downward orientation is coupled with a negative one. This suggests that target concepts in orientational metaphors are structured in a coherent and uniform manner (Kövecses, 2010). Equivalent examples can be provided for center - periphery, left - right, front – back, near – far, and other orientations.

Ontological metaphors “involve the projection of entity or substance status on something that does not have that status inherently” (Lakoff & Johnson, 1980a: 196). In other words, “ontological metaphors enable us to see more sharply delineated structure where there is very little or none” (Kövecses, 2010: 39). This means that ontological metaphors such as INFLATION IS AN ENTITY, THE MIND IS A MACHINE, THE MIND IS A BRITTLE

OBJECT, THE MIND IS A CONTAINER, COUNTRY IS A CONTAINER, RACE IS A CONTAINER, THEORY IS A LIVING ENTITY, and

INFLATION IS A LIVING ENTITY (Lakoff & Johnson, 2003[1980b]: 25-34), enable us to make sense of some more abstract, intangible concepts, by relying on our experiences with physical objects. This is particularly transparent in the case of the CONTAINER metaphor, which is largely based on the experience of our bodies as containers. 2.1.2. Systematicity and Partial Nature of Metaphorical Mappings

The fact that a concept is structured by a metaphor means “it is partially structured and that it can be

extended in some ways but not in others” (Lakoff & Johnson, 2003[1980b]: 13), and, in addition, such


“metaphorical structuring of a concept is necessarily partial” (ibid.: 52). This, coupled with the fact that “concepts are metaphorically structured in a systematic way” (ibid.: 52), means that expressions from the source domain can be used to refer to concepts in the target domain. Additionally, due to its systematicity, “features of the source and target domain are joined so that the metaphor may be extended or have its internal logic” (Saeed, 2003: 348). Let us consider the metaphor ARGUMENT IS WAR, discussed in Lakoff and Johnson (2003[1980b]: 4-6). Namely, conceptualizing arguments in terms of war sanctions the use of vocabulary such as attack, defense, victory, and defeat, to talk about arguments. The systematicity of this metaphor allows us to talk about interlocutors as opponents who can win an argument, succumb to the attacks, or lose a battle. Still, such systematicity is necessarily paired with the partial nature of metaphorical mappings, suggesting that not everything is projected from the source to the target. In other words, the partial nature of metaphorical mappings necessarily involves a certain degree of highlighting and hiding (Lakoff & Johnson (2003[1980b]; Kövecses, 2010). 2.1.3. Metaphorical Entailments

Another important element in the CMT framework includes metaphorical entailments, which refer to rich additional knowledge structures about the source domain that are transferred to the target domain in the process of cross-domain mapping (Evans & Green, 2006; Kövecses, 2010). In other words, “since metaphorical concepts are defined in terms of nonmetaphorical concepts, they show entailment relations parallel to those for the corresponding nonmetaphorical concepts” (Lakoff & Johnson, 1980a: 197). Owing to the fact that “metaphoric mappings carry entailments or rich inferences” (Evans & Green, 2006: 298), it is possible to use additional knowledge about the source domain in order to make sense of some possible features of the target domain (Kövecses, 2010). As suggested by Evans and Green (2006: 299), and Kövecses (2010: 122), beside the conventional mappings for the metaphor ARGUMENT IS A JOURNEY, some additional knowledge can also be inferred: travelers can stray from the path, be forced to take a break, get lost, not reach their destination, etc. Consequently, interlocutors can stray from the subject, be forced to abandon the discussion, be unable to make their point, or be unable to reach an agreement. This suggests that the association that exists between the source and the target domain makes it possible for the events that normally occur in the source to occur in the target as well. 2.1.4. Metaphor Systems and Metaphorical Coherence

An important aspect of metaphors is that they can interact with each other and yield elaborate metaphor systems constructed out of individual conceptual metaphors (Lakoff & Johnson, 2003[1980b]; Evans & Green, 2006). Metaphor systems represent groups of conceptual metaphors that offer partial structurings of


a concept, and they can be understood as instantiations of one general conceptual metaphor situated at the highest point in the hierarchy, where ”metaphors higher up in the hierarchy tend to be more widespread than those mappings at lower levels” (Lakoff, 2006[1993]: 209-210). One of the metaphor systems discussed in Kövecses (2010: 149-167) includes the Event Structure Metaphor which accounts for the metaphorical understanding of events.

Namely, Lakoff (2006[1993]) found that different aspects of event structure are defined in metaphorical terms, and consequently proposed the general mapping for the Event Structure Metaphor:

− States are locations (bounded regions in space). − Changes are movements (into or out of bounded regions). − Causes are forces. − Actions are self-propelled movements. − Purposes are destinations. − Means are paths (to destinations). − Difficulties are impediments to motion. − Expected progress is a travel schedule; a schedule is a virtual traveler, who reaches prearranged destinations at

prearranged times. − External events are large, moving objects. − Long term, purposeful activities are journeys. (Lakoff, 2006[1993]: 204)

These individual submappings are understood to constitute the more general Event Structure Metaphor. In addition, each submapping maintains its entailments. For example, Changes are Movements entails that “lack of control over change is viewed as lack of control over movement” (Kövecses, 2010: 164), and “that accidental changes are conceptualized as accidental movements” (ibid.: 164). Metaphors LIFE IS A JOURNEY and ARGUMENT IS A JOURNEY are also said to inherit their structure from the more general Event Structure Metaphor (Evans & Green, 2006). In this sense, the existence of mappings lower in the hierarchy is sanctioned by the general metaphor.

Various metaphors used to describe the concept of argument also emphasize different aspects of that concept. The involvement of source domains of WAR, JOURNEY, CONTAINER, and BUILDING is possible owing to the overlapping of purposes that underlie their use. Additionally, this overlapping of purposes also means that these metaphors are coherent. For example, Lakoff and Johnson (2003[1980b]: 94) suggest that metaphorical coherence between the conceptual metaphors ARGUMENT IS A JOURNEY and ARGUMENT IS A CONTAINER arises from the overlap of entailments between them, which also leads to “a partial satisfaction of their purposes” (ibid.: 95). Namely, the overlaps in purposes underlying the use of specific metaphors that correspond to the overlaps in metaphors “can be characterized in terms of shared metaphorical entailments” (ibid.: 97). Specifically, the overlap between these two metaphors resides in the fact that both journeys and containers define surfaces. This represents the case of coherence between two aspects of a single concept.

When it comes to complex coherence across metaphors, Lakoff and Johnson (2003[1980b]: 97-105) discussed the following metaphors: ARGUMENT IS A JOURNEY, ARGUMENT IS A CONTAINER, and ARGUMENT IS A

BUILDING. The authors suggest that the coherence between these metaphors is again “based on the fact that all three have content-defining surfaces” (ibid.: 102). However, content is defined differently by a surface in


each of the cases. Namely, in JOURNEY metaphors, content is defined by the ground covered by an argument; in CONTAINER metaphors, the content is limited by the surface of the container; in BUILDING metaphors the content is not inside the building, but in its foundation and the outer shell. In all, these three metaphors “are part of whole metaphorical systems that together serve the complex purpose of characterizing the concept of an argument in all of its aspects” (ibid.: 105). Owing to the overlap in their entailments, they are also coherent. 2.1.5. Image Schemas and the Reality behind Them

In line with the notion of embodied cognition and the emphasis on the importance of categorization, cognitive linguists have introduced the notion of image schemas, where “an image schema is a condensed redescription of perceptual experience for the purpose of mapping spatial structure onto conceptual structure” (Oakley, 2007: 215). In other words, image schemas represent preconceptual structures derived from everyday patterns of bodily interaction that provide the scaffolding for the complex conceptual systems that govern human cognition. Lakoff (1990) tackled the issue of the relationship between abstract reasoning and image schemas, and suggested that “a great many, if not all abstract inferences are actually metaphorical versions of spatial inferences that are inherent in the topological structure of image schemas” (ibid.: 54). In addition, image schemas “can be extended by a process of metaphorical extension into abstract domains” (Saeed, 2003: 355), such as politics.

Johnson (1987: 126, cited in Oakley, 2007: 217) offered the following list of basic image schemas: CONTAINER, BALANCE, COMPULSION, BLOCKAGE, COUNTERFORCE, RESTRAINT REMOVAL, ENABLEMENT, ATTRACTION, MASS-COUNT, PATH, LINK, CENTER-PERIPHERY, CYCLE, NEAR-FAR, SCALE, PART-WHOLE, MERGING, SPLITTING, FULL-EMPTY, MATCHING, SUPERIMPOSITION, ITERATION, CONTACT, PROCESS, SURFACE, OBJECT, COLLECTION. In addition, image schemas can undergo transformations that are also very important, since human cognition is grounded in the notion that allows us to map the perceptual experience licensed by image schemas onto more abstract conceptual categories (Oakley, 2007; Gibbs & Colston, 2006[1995]). The primary transformations include (ibid.):

− Path focus to end-point focus (to understand this transformation, follow an imaginary path of an object, and then focus on the point where the object comes to rest, or where it will come to rest).

− Multiplex to mass (imagine a group of individual objects, and then start moving away from them. As you are doing this, note how the group is slowly transformed into a homogenous mass. As you begin to move back towards them, the mass is transformed into individual objects).

− Trajectory (if we observe a moving object, we can mentally trace its path). − Superimposition (if we imagine a large sphere and a small cube, we can mentally enlarge the size of the

cube so that it can accommodate the sphere; we can then reduce the size of the cube so that it fits the inside of the sphere).

The role of image schemas has also been emphasized in the developmental context (e.g. Mandler,

1992; 2012), where it is claimed that “image schemas provide the earliest meaning available to the infant for purposes of preverbal thought ... and form the basis on which natural language rests” (Mandler, 1992: 592).


This in turn prompted the author to label image schemas as conceptual primitives, where she defines them as “dynamic analog representations of spatial relations and movements in space” (ibid.: 591), where “these new representations are the primitive meaning elements used to form accessible concepts” (ibid.: 591). In addition, Mandler (2012) gives an in-depth analysis of the reasons why the conceptual system is grounded in spatial experience. Namely, the author presents the following evidence to support her claims: spatial movement is very salient and can easily attract attention; it is easily understandable; a fully operational conceptual system for making sense of objects and events can be formed based on spatial representations alone; spatial representations have a clear structure that renders them memorable; spatial representations can be imaged, which means that they can be recalled.

Gibbs and Colston (2006[1995]) presented evidence for the cognitive psychological reality of image schemas and their transformations, based on the research results from the fields of psycholinguistics, cognitive psychology, and developmental psychology. In the psycholinguistic context, the authors discussed research conducted by Gibbs et al. (1994, cited in Gibbs & Colston, 2006[1995]: 243-246), where it was found that “people make sense of different uses of stand because of their tacit understanding of several image schemas that arise partly from the ordinary bodily experience of standing” (Gibbs & Colston, 2006[1995]: 245). One of the most interesting findings in the domain of cognitive psychology pertains to the visual and auditory representational momentum that is also explained in terms of image schemas and their transformations (see Gibbs & Colston, 2006[1995]: 248-253 for details). One of the examples in the area of developmental psychology includes a study conducted by Wagner et al. (1981, cited in Gibbs & Colston, 2006[1995]: 259), where it was found that nine to 12-month-old infants spent more time looking at the dotted line than the solid line when a pulsating tone was played, which suggested that the infants were able to construct a metaphorical match. Also, when listening to a rising tone, the infants looked more at the arrow pointing upward, whereas when listening to a descending tone they spent more time looking at the arrow pointing downward. All of these findings suggest that image schemas are not fictional derivatives, but actual constructs that not only enable us to make sense of the world, but are also monumental for our ability to establish connections between more abstract domains of experience. 2.1.6. The Invariance Principle

An important aspect of the CMT includes restrictions that constrain certain cross-domain mappings, while at the same time licensing others. For example, in the case of generic-level metaphors, Lakoff and Johnson (1989, cited in Lakoff, 1990: 68-69) found that death was personified in a rather small number of linguistic realizations. The investigation of such constraints first gave way to the Invariance Hypothesis (Lakoff, 1990) that later evolved into the Invariance Principle (Lakoff, 2006[1993]):

“Metaphorical mappings preserve the cognitive topology (this is, the image-schema structure) of the source domain, in a way consistent with the inherent structure of the target domain” (ibid.: 199).


Obviously, the Invariance Principle is largely grounded in the notion that stresses the importance of

image-schematic structures in human cognition. In plain terms, the metaphorical cross-domain mappings make the source and the target domain structures consistent with each other, explaining why some pairs of source-target domains are facilitated, while others are restricted. When viewed at the image-schematic level, the Invariance Principle suggests that in the case of container schemas “interiors will be mapped onto interiors, exteriors onto exteriors, and boundaries onto boundaries” (ibid.: 199). Similar conclusions can be derived for other types of schemas.

In addition, the Invariance Principle also implies that a metaphor will allow the transfer of only those metaphorical entailments that are compatible with the target domain. This is labeled as the target domain

override and it represents the process “that prevents entailments from projecting to the target domain” (Evans & Green, 2006: 303), which is “a consequence of the fact that inherent target domain structure automatically limits what can be mapped” (Lakoff, 2006[1993]: 200). In short, metaphorical mappings are constructed in such a fashion that preserves the basic topology of the target domain, i.e. the image schematic structure of the target domain is invariant. 2.1.7. Frames and ICMs

Charles Fillmore introduced the notion of frames, where a frame refers to “any system of concepts related in such a way that to understand any one of them you have to understand the whole structure in which it fits” (Fillmore, 2006[1982]: 373), or, in other words, “a frame is a structured mental representation of a conceptual category” (Kövecses, 2006: 64). In this sense, the term frame is intimately related to word meaning, and in Fillmore’s frame semantics “words represent categorization of experience” (Fillmore, 2006[1982]: 373). This promotes the view according to which word meaning is largely conditioned by the perceptual experience and is related to a set of prototypes. In addition to prototypicality effects, the frame against which a word meaning is understood is also grounded in cultural experience (Kövecses, 2006).

For example, let us consider the COMMERCIAL EVENT frame as discussed in Fillmore (2006[1982]: 378). Namely, this frame includes the following elements: the buyer, the seller, the goods, and the money. Furthermore, this frame can be invoked through the use of various verbs, and specific verbs can even be used to place focus on certain elements of the frame, while backgrounding others, which is reminiscent of Lakoff and Johnson’s (2003[1980b]) notion of highlighting and hiding. The verb buy, for example, places the focus on the buyer and the goods, the verb sell places the focus on the seller and the goods, etc. This suggests that understanding the meaning of these verbs requires the recruitment of the COMMERCIAL EVENT frame. In short, “a speaker ‘applies’ a frame to a situation, and shows that he intends this frame to be applied by using words recognized as grounded in such a frame” (ibid.: 382).


Based on the research conducted by Fillmore, Kövecses (2006: 67-69) presented some additional characteristics of frames: 1) frames are evoked by particular meanings of words; 2) focus can be placed on particular elements of a frame; 3) frames can be used to impose a perspective on a situation; 4) frames provide a particular history; 5) some frames incorporate larger cultural frames; 6) frames are idealizations. In turn, all of the mentioned characteristics contribute to the view that “frames constitute a huge and complex system of knowledge about the world” (ibid.: 69).

Building on Eleanor Rosch’s Prototype Theory, George Lakoff (1987, cited in Evans & Green, 2006: 269-281) introduced the notion of Idealized Cognitive Models (ICMs) to account for prototypicality effects. In this sense, ICMs represent “relatively stable knowledge structures that are built up on the basis of repeated experience” (Evans & Green, 2006: 279), and “they are ‘idealized’ because they abstract across a range of experiences rather than representing specific instances of a given experience” (ibid.: 270). In plain terms, ICMs are recruited during online processing in order to help us make sense of the conversation. Owing to their connection to typicality effects, ICMs stand as an essential tool for categorizing the world around us. Additionally, ICMs are not innate, but rather emergent structures that arise from experience and everyday interaction with the world and objects in it. Their connection to various levels of prototypicality renders them idealized, in that they are connected to the most general and most abstract representations. In other words, “the models are idealized, in that they involve an abstraction, through perceptual and conceptual processes, from the complexities of the physical world” (Cienki, 2007: 176). In addition, the degree of prototypicality is often connected to the level of conventionality, suggesting the existence of a cultural basis, similar to that of frames.

In summary, what both frames and ICMs have in common is the fact that “they all designate a coherent organization of human experience” (Kövecses, 2006: 64). 2.1.8. Conventional and Novel Metaphors

Charteris-Black (2004) emphasizes the distinction between conventional and novel (creative)

metaphors. Namely, conventional metaphors “reflect a diachronic process whereby use that was originally ‘metaphorical’ becomes established as ‘literal’ within a language” (Charteris-Black, 2004: 17). In other words, conventional metaphors are licensed by the entrenched conceptual associations. Additionally, a novel metaphor can easily become conventionalized through an increased frequency of use, which serves as the main marker of the degree of metaphor conventionality when analyzing language corpora. However, although a conventional metaphor may not stand out in discourse in a way that a novel metaphor might, this in no way undermines its value and potential. In fact, its ability to go on unnoticed renders it a potent tool for perception management and manipulation, especially in political discourse.


In terms of differences in interpretation between conventional and novel metaphors, Coulson and Oakley (2005: 1524) argue that conventional, i.e. entrenched metaphors, involve an automatic process of retrieval, whereas novel metaphors require additional computations and analogical reasoning. This means that the emergent structure1 in novel metaphors needs to be actively constructed, while in the case of entrenched metaphors it is already available for recruitment.

Another important point that needs to be stressed concerns the distinction between a conceptual

metaphor and a metaphorical expression. Namely, the term ‘metaphor’ is used “to refer to the conceptual mapping, and the term ’metaphorical expression’ to refer to an individual linguistic expression ... that is sanctioned by a mapping” (Lakoff, 2006[1993]: 192). In other words, “metaphorical expressions are the linguistic manifestation of underlying conceptual knowledge” (Coulson, 2006b: 33), which reflects the idea that “metaphor is not just a matter of language, but of thought and reason” (Lakoff, 1990: 49), where language is seen as “a reflection of the mapping” (ibid.: 49). In this sense, a conventional metaphor can be understood as an entrenched trigger that fires out individual metaphorical expressions.

2.1.9. Critical Metaphor Analysis (CMA)

Given the ubiquitous nature of metaphors in everyday life, recognizing them and understanding the factors that underlie their use in various types of discourse is essential. This necessity becomes even more transparent when the potential for ideological manipulation typical of political discourse is taken into consideration. Charteris-Black (2004: 243-253) outlines the basic tenets of Critical Metaphor Analysis (CMA), and stresses that “an awareness of their [metaphors’] motivation in socially influential domains of language use improves our understanding of the ideological basis for metaphor choice” (ibid.: 244). Bearing in mind that the process of cross-domain metaphorical mappings goes on at an unconscious level, it is clear why CMA can serve as a potent tool for unraveling the hidden persuasive functions sanctioned by metaphors.

Figure 1. A discourse model for metaphor (adopted from Charteris-Black, 2004: 248)

1 The notion of emergent structure is described in section 2.2.2 below.

Metaphor and Politics 20

In his approach, Charteris-Black (2004: 244) offers the following conceptual levels for metaphor analysis that are hierarchically organized: conceptual keys, conceptual metaphors, and metaphors. Conceptual keys are situated at the highest point in the hierarchy, followed by conceptual metaphors, and metaphors that stand for metaphorical expressions. According to this author, the choice of metaphors in discourse is governed by their persuasive function, which is obvious in his discourse model of metaphor presented in Figure 1. As shown in the diagram, various factors conspire in order for the persuasive rhetorical function of a metaphor to be achieved. Consequently, “understanding more about metaphor is an essential component in intellectual freedom” (ibid.: 253). 2.1.10. Metaphor and Politics

“The very systematicity that allows us to comprehend one aspect of a concept in terms of another ... will necessarily hide other aspects of the concept” (Lakoff & Johnson, 2003[1980]: 10), which means that a metaphor can make us see only certain aspects of an issue, while rendering us blind to its other sides. Also, “owing to their ability to transform the abstract onto the concrete, metaphors represent one of the most significant instruments of persuasion and propaganda in the language of political rhetoric”2 (Radić-Bojanić & Silaški, 2008: 141), suggesting that “political metaphors can justify courses of action” (Mio, 1997: 118), and can be used to “create political reality” (ibid.: 121). In addition, conceptual metaphors can “make the public feel a part of the political process and supportive of decisions made by the political elite” (ibid.: 118), rendering their value systems more susceptible to manipulation. The choice of such metaphors is often not arbitrary, and politicians use them consciously, since “the fact that by using metaphors they choose which aspects of a concept will be emphasized, creators of a text more or less consciously reveal their value judgments and ideological stances.”3 (Silaški, Đurović, & Radić-Bojanić, 2009: 119).

According to Mio (1997), the three basic mechanisms that enable a conceptual metaphor to perform its persuasive function in political discourse include: 1) simplification; 2) manipulation of underlying symbolic representations; and 3) the emotional appeal. Firstly, if a political issue is too complex or abstract for an average supporter to grasp, a metaphor can be used to simplify it. For example, the conceptual metaphor POLITICS IS SPORT can transform the complexity of a political issue into a simple contest. Secondly, after enough repetitions metaphorical connections can become so entrenched as to go on unnoticed. In other words, “what was initially understood to be metaphorical can become implicit and accepted as truth at an unconscious level” (ibid.: 122). In that sense, conventional, i.e. entrenched metaphors can serve as a better conduit for political ideas compared to novel, i.e. creative metaphors, precisely due to the fact that the former

2 „Zbog svoje sposobnosti prenošenja apstraktnog na konkretno, metafore predstavljaju jedan od najvažnijih instrumenata ubeđivanja i propagande u jeziku političke retorike“. 3 ,,Time što putem metafora biraju koje će aspekte nekog pojma istaći, tvorci teksta manje ili više svesno otkrivaju svoje vrednosne sudove i ideološke stavove“.


can go on undetected (Burnes, 2011). Finally, the role of metaphors in stirring emotions is invaluable, since “when emotions are evoked, logic is circumvented” (Mio, 1997: 123) and persuasion more easily achieved.

The persuasive aspect of metaphors in political discourse and their emotional appeal are also addressed in Mio (1996). His experimental research provided empirical support for the Metaphor Extension

Hypothesis, according to which “metaphors that extend someone else’s metaphor are more effective persuasive devices than those that do not” (ibid.: 136). To arrive at such a conclusion, the author investigated political debates where one of the participants constructed his arguments around the metaphor(s) introduced by his opponent. Such an approach was rated as more persuasive compared to the cases where participants introduced new metaphors. Additionally, research performed by Mio et al. (2005) showed that US presidents who used more metaphors in their inaugural speeches were labeled as more charismatic than those who used fever metaphors. Also, “even those presidents who did not appear to be charismatic were still perceived to be more inspiring when they used metaphors“ (ibid.: 292), which suggests that metaphor really is a powerful rhetorical tool.

Charteris-Black (2004; 2009; 2011) also stressed the importance of emotional dimensions that metaphors introduce into the architecture of political rhetoric. Namely, building on the Aristotelian notion of rhetoric and aspects of contemporary rhetoric, Charteris-Black (2009; 2011) discusses the notions of ethos (a morally worthy stance), logos (proofs that support an argument), and pathos (arousing emotions). Their relationship is illustrated in Figure 2 below, where it is obvious that all of these rhetorical elements are connected by their persuasive aspect.

Figure 2. Rhetorical means of persuasion in political communication (adopted from Charteris-Black, 2004: 14)

Charteris-Black (2009; 2011) proposes metaphor as one of the main tools in political persuasion, and

presents its ability to arouse emotions as one of the key mechanisms in persuasion. Namely, “metaphors change how we understand and think about politics by influencing our feelings and thoughts” (Charteris-Black, 2011: 32), which suggests that “increasing the emotional impact is a very vital role for metaphor” (Charteris-Black, 2009: 104). In other words, “the social role of metaphor in the construction of an ideology is motivated by a rhetorical purpose of arousing the emotions in order to persuade” (Charteris-Black, 2004: 251). In addition to heightening the pathos, Charteris-Black (2011) also claims that politicians often resort to metaphors in order


to justify their arguments and present them as rational. In this sense, political metaphors are understood to advocate only the desired stances affiliated to particular groups (Charteris-Black, 2009), which is reminiscent of Lakoff and Johnson’s (2003[1980b]) notion of highlighting and hiding. Additionally, metaphors are often used by politicians either for self-representation, or to evaluate others and their actions.

Thompson (1996) emphasizes the fact that politics is grounded in the group-level phenomena, suggesting that “a major function of political metaphors is to link the individual and the political” (ibid.: 186), and that “the existence of metaphors describing the links between individuals and larger groups is crucial to making politics relevant” (ibid.: 187). In addition, the author claims that metaphors facilitate communication between the political elites and their supporters, which is sanctioned by the fact that metaphors combine values, emotions, and ideology. Furthermore, “metaphors have implications for action because of their ability to frame issues” (ibid.: 194), suggesting that apart from being grounded in our physical and social experience, “they also influence our experience and our actions” (Lakoff & Johnson, 2003[1980b]: 68).

Let us consider the following example: “When the Lord asked the prophet Isaiah, he said, ‘Whom shall I send and who will go for us?’ Isaiah said, ‘Here am I, send me,’” Mr. Perry said in Council Bluffs. “This is your country. Taking her back is your calling. Join me in this mission.” (28 December 2011, The New York Times online edition)

The above example reveals the underlying conceptual metaphor POLITICS IS RELIGION. This metaphor presents a political campaign in terms of a holy mission, rendering the politician a saint fighting for a holy cause. Naive as it may appear at first, this simple metaphor has a very powerful emotional appeal with the public. Namely, it strikes upon the very place where some of them are most sensitive - their religious beliefs. The frame of RELIGION also encompasses a whole set of values, and by appealing to those values, the politician is practically blackmailing his supporters into following him, because not doing so would be a sin that would interfere with their established value system. This is a transparent example of how “politicians commonly reframe the ideas expressed in relatively neutral language; that is, they place ideas in frames that were never intended to be used in conceptualizing ideas” (Kövecses, 2006: 92), which gives way to silent manipulation. In other words, “the manner in which one frames an issue can have practical implications for social policy and for politics” (Cienki, 2007: 174). Additionally, “we conceptualize experience by means of frames we get from a particular culture, and these frames are anything but neutral” (Kövecses, 2006: 91), suggesting that the (ab)use of metaphorical language in political discourse can give way to the creation and implementation of new frames, and therefore entirely new sets of values. Now it is becoming even more evident why “metaphors are regarded as a powerful persuasive tool used for generating or exploiting some of the already existent4 specific value judgments, thus constructing via discourse a certain point of view” (Silaški & Đurović, 2011: 229).

The lexical cue mission can also serve to introduce the JOURNEY metaphor, and the even more elaborate EVENT STRUCTURE metaphor. In this sense, the previous example can be said to introduce the following metaphor system: POLITICS IS RELIGION → RELIGION IS A JOURNEY → POLITICS IS A JOURNEY (where

4 My italics.


journey is understood in terms of event structure). Such a system can serve to obscure the persuasive aspects of a metaphor even further, owing to its coherence and the shared entailments. Also, the JOURNEY metaphor is derived from an even more fundamental image schema of PATH, lending support to Lakoff’s hypothesis that abstract reason is based on image schemas (Lakoff, 1990). With all of this in mind, it is safe to state that metaphor “conceals an underlying persuasive function that is often not immediately transparent” (Charteris-Black, 2004: 9), by manipulating the fundamental conceptual structure of the human cognitive system. This in turn can be used to silently shape the collective consciousness of the public and pave the way for ideologies that represent the building blocks of entire belief and value systems. 2.1.11. Section Summary

In brief, the current section introduced some of the basic tenets of the CMT that will be used in the present research. Additionally, the notions of image schemas, frames, and ICMs have also been discussed in some detail, owing to the fact that these constructs will be used in the analyses. Finally, the basic framework of the Critical Metaphor Analysis (Charteris-Black, 2004) was introduced, which was followed by the discussion of the role of conceptual metaphors in political discourse. Additionally, some of the basic mechanisms via which metaphors achieve their persuasive function in political discourse have also been discussed, with the ability to provoke an emotional appeal presenting itself as one of the more salient of these mechanisms.

Conceptual Blending Theory 24

2.2. Conceptual Blending Theory (CBT)

This section of the paper will introduce the basic tenets of the Conceptual Blending Theory, best elaborated in Gilles Fauconnier and Mark Turner’s seminal book The Way We Think: Conceptual Blending and

the Mind’s Hidden Complexities. The section will start by addressing the notion of mental spaces and the Mental Space Theory as introduced by Fauconnier (1994; 1997; 2007), with the CBT as its logical ‘sequel’. In addition, special attention will be paid to the treatment of metaphor within the CBT framework (e.g. Grady et al., 2007[1999]; Fauconnier & Turner, 2008; Coulson, 2006b), and the section will close by providing a comprehensive comparison between the CMT and the CBT, advocating for the possibility of a joint use of the two frameworks.

2.2.1. Mental Spaces and Backstage Cognition

The onset of the 2nd generation cognitive science marked a shift from the traditional truth-conditional notion of meaning, highlighting the idea of meaning construction as a dynamic process largely conditioned by the context and background knowledge structures such as frames and cognitive models (Fauconnier, 1994; 1997; 2007; Evans & Green, 2006). Although it is understood that language essentially does encode meaning, linguistic structures introduced into the discourse serve as partial prompts for meaning construction, the process often labeled as conceptualization. The process of conceptualization is licensed by the human cognitive architecture that incorporates entire sets of unconscious actions dubbed “backstage cognition” (Fauconnier, 1994; Fauconnier & Turner, 2002). In an effort to develop these ideas further, Gilles Fauconnier introduced the notion of mental spaces which represent “domains that discourse builds up to provide a cognitive substrate for reasoning and for interfacing with the world” (Fauconnier, 1994: 34). In other words, mental spaces are “very partial assemblies constructed as we think and talk for purposes of local understanding and action” (Fauconnier, 2007: 351). According to the Mental Space Theory, the two main stages in meaning construction include: 1) creation of a mental space; and 2) formation of mappings between spaces, where mappings are influenced by the local discourse context (Evans & Green, 2006).

Mental spaces are normally set up by space builders, where “a space builder is a grammatical expression that either opens a new space or shifts focus to an existing space” (Fauconnier, 1997: 40). Some typical space builders include “prepositional phrases, adverbials, subject-verb complexes, conjunctions + clause” (Fauconnier, 2007: 371), etc. The topology of a mental space is structured by frames or ICMs, and elements that constitute mental spaces are either pre-existent and introduced by frames (a process often labeled as schema induction, Evans & Green, 2006), or are constructed on-line. Elements are introduced via linguistic expressions, noun phrases (NPs), that can either have a definite or an indefinite interpretation, which is marked grammatically by the presence of a definite, or an indefinite article, respectively (Fauconnier, 1994).


Consequently, NPs with an indefinite interpretation are used to introduce novel elements, whereas “NPs with definite interpretation are said to function in the presuppositional mode, because they presuppose existing knowledge” (Evans & Green, 2006: 372), but they still need not necessarily have a unique referent. Alternatively, new elements can also be introduced by “nonlinguistic pragmatic conditions (e.g. objects which are salient in the interaction that produces the discourse)” (Fauconnier, 1997: 39).

Additionally, “definite descriptions are primarily role functions and secondarily the values taken by such roles” (Fauconnier, 1994: 40). In that sense “mental spaces ... belong to the domains of role functions” (ibid.: 42), which suggests that roles can adopt different values in different spaces. In broad terms, roles are introduced by frames. For instance, the election frame introduces, among others, the role of the president, and the competing political parties and their representatives. Once this frame has been recruited via the process of schema induction, those roles are connected to their values, i.e. specific participants in the elections who vary from country to country, or over longer time periods. In other words, the abstract role-scaffolding introduced by the frame is completed by the context-bound values that can vary across spaces.

As discourse unfolds, new mental spaces are introduced, creating structures called lattices (Fauconnier, 1994; 1997; 2007). An important aspect of these networks are various mappings that serve to connect elements across spaces, with identity and analogy being the most common (Fauconnier, 2007). A crucial theoretical construct that facilitates this process is the Access Principle 5 which states:

“If two elements a and b are linked by a connector F (b = F(a)), then element b can be identified by naming, describing, or pointing to its counterpart a.” (Fauconnier, 1997: 41)

Element a is labeled as the trigger, while b is the target. This means that the description of the trigger in one space can be used to access the target in another space. Additionally, owing to the Access Principle, “expressions referring to a particular counterpart can typically provide access to entities in mental spaces in either direction” (Evans & Green, 2006: 377), suggesting a bidirectional nature of the mappings, typical of open connectors.

In order for interlocutors to be able to navigate through the discourse lattice, it is necessary for them to keep track of the base, viewpoint, and focus (Fauconnier, 1997). The base space is “the starting point for the construction to which it is always possible to return” (ibid.: 49); the viewpoint represents the space “from which discourse is currently being viewed and from which other spaces are currently being built” (Evans & Green, 2006: 389); the focus space represents the space that is currently under construction. In light of the tense-

aspect-mood system6 in Mental Space Theory, another important space in the lattice is the event space, which “represents the time associated with the event being described” (ibid.: 389). For instance, tense-marking which is associated with the time in the event space can shift focus from one space to another.

5 The Access Principle was initially introduced as the Identification Principle (for details see Fauconnier, 1994: 3). 6 For a detailed account of the tense-aspect-mood system in Mental Space Theory see Fauconnier, 1997: 72-98; Fauconnier, 2007: 365- 371; Evans & Green, 2006: 386-396.


When a space builder introduces a new mental space in discourse, that new space will be included in its parent space (Fauconnier, 1994), suggesting a hierarchy in the organization of mental spaces. Consequently, “there must be a connector capable of connecting triggers and targets in the parent and daughter spaces” (ibid.: 18). Additionally, “structure from a parent space is transferred to a new space by default ... the default transfer, called optimization, will apply to the extent that it does not contradict explicit structure in the new space” (Fauconnier, 2007: 355-356), which in turn constitutes the Optimization Principle.

With some of the basic tenets of the Mental Space Theory explained, the following sections will pursue its logical ‘sequel’ – the Conceptual Blending Theory.

2.2.2. Conceptual Blending Conceptual blending, alternatively discussed as the many space model (Turner & Fauconnier, 1995) or the network model of conceptual integration (Fauconnier & Turner, 2006[1998]), was developed by Gilles Fauconnier and Mark Turner as a theory of online meaning construction. In short, the paradigm “is concerned with on-line, dynamical cognitive work that people do to construct meaning for local purposes of thought and action” (Fauconnier & Turner, 2006[1998]: 312), while “a fundamental motivating factor of blending is the integration of several events into a single unit” (ibid.: 332). As a ubiquitous phenomenon, blending is understood to be the driving force not only behind human imagination, but also behind some basic cognitive processes that constitute our everyday realities. In other words, “blending is pervasive in all modes of thinking and talking, and in fact it is not even inherently tied to language, but appears more generally in action and phenomena of cognition” (Fauconnier & Turner, 1994: 7).

The so-called “minimal network” (Fauconnier & Turner, 2006[1998]: 313; 2002: 47) in conceptual blending consists of four spaces: the generic space, two input spaces, and the blended space or the blend (Fauconnier & Turner, 2006[1998]; 2002; Coulson & Oakley, 2000; Turner, 2007). Alternatively, the blending network can include multiple input spaces, or even multiple blends. The generic space maps onto each of the input spaces, and it “contains what those inputs have in common at any moment in the development of the conceptual integration network” (Fauconnier & Turner, 2006[1998]: 308). In that sense, it “contains highly schematic information which serves as a basis for establishing cross-space mappings between the input spaces” (Evans & Green, 2006: 406). These cross-space mappings facilitated by the generic space serve to identify counterparts in the inputs, which are then projected to the blend. Input spaces represent partial structures that normally correspond to events that are to be integrated in the blended space. Inputs are involved in selective projections which means that “only the matched information, which is required for purposes of local understanding” (ibid.: 409) is projected to the blend. Such partial nature of the projections can yield the production of different blends, even though the inputs remain the same. The blended space appears as the result of conceptual integration, and its main element is the emergent structure not present in


either of the inputs (Fauconnier & Turner, 2006[1998]; 2002; Coulson & Oakley, 2000). The emergent structure in the blend can cause backward projections7 to the inputs (Fauconnier & Turner, 2006[1998]), which can influence global insight at the human scale.8 Such backward projections are licensed by the Access Principle.9

Emergent structure in the blend appears as the result of the following processes: composition, completion, and elaboration. Composition includes “attributing a relation from one mental space to an element or elements from other input spaces” (Coulson & Oakley, 2000: 180), and in addition, elements from the inputs can be composed in such a way, as to create relations that are not present in either of the inputs (Turner, 2007; Fauconnier & Turner, 2006[1998]). As a result, emergent structure is a product of “contextual accommodation of a concept from one domain to apply to elements in a different domain” (Coulson & Oakley, 2000: 180). The process of completion is based on the fact that a lot of background knowledge and conceptual structure gets recruited into the blend (Turner, 2007; Fauconnier & Turner, 2006[1998]; 2002), which means that the above described “composed structure is completed with other structure” (Fauconnier & Turner, 2006[1998]: 314). In other words, “completion involves schema induction which refers to 10

unconscious and effortless recruitment of background frames” (Evans & Green, 2006: 409). A transparent form of completion is pattern completion, where, for example, an event can be completed by information based on background knowledge or preexistent entrenched frames stored in long-term memory (Fauconnier & Turner, 2006[1998]; Coulson & Oakley, 2000; 2006). Elaboration or “running of the blend” (Fauconnier & Turner, 2002: 44), involves mental simulations “according to principles and logic in the blend” (Fauconnier & Turner, 2006[1998]: 315). In other words, elaboration represents “the simulated mental performance of the event in the blend, which may continue indefinitely” (Grady, Oakley & Coulson, 2007[1999]: 425). Elaboration can appear in two forms: coupled and decoupled (Coulson & Oakley, 2000). Decoupled elaboration “involves little or no physical realization” (ibid.: 181), whereas coupled elaboration involves physical realization, i.e. actual physical activity. Coulson and Oakley (2000) suggest that the latter type of elaboration can include action blends, with patterns of activity from one domain applied to elements of another domain. 2.2.3. The Generic Space Problem

Based on the current research trends in the domain of the CBT, the role of the generic space and even its psychological plausibility have been largely disputed (Brandt & Brandt, 2005; Oakley, 2011; Coulson & Oakley, 2005). Namely, the structures present in the generic space are “immanent in the input and blended spaces themselves and are of little relevance to the network as a cognitive model of online meaning

7 Note that Coulson (2001: 178-196) refers to backward projections as retrospective projections. 8 For more details concerning the notion of human scale consult subsection 2.2.5. 9 See section 2.2.1 above. 10 My italics.


construction” (Oakley, 2011: 4). One of the possible issues where a generic space could play an important role is in “modeling the socio-historical development of ideas over larger temporal spans” (ibid.: 4), such as complex numbers. In that sense, the present research will adopt the stance that the generic space is largely an artifact of analysis, and shall treat it as such. However, it will be argued that the structure of this “artifact” can be represented at the image schematic level. Namely, in order to avoid the generativist pitfall and its stipulation that, owing to the fact that the generic space contains structure common to the entire network, the generic space can be viewed as a generative mechanism largely similar to the Chomskian notion of the LAD, the present research will instead argue that the generic space, although largely an artifact of analysis, can be represented in the form of the image schematic structure that provides the scaffolding on top of which the human ability for meaning construction is mounted.

Building on the notion of image schemas as conceptual primitives (Mandler, 1992), and the discussion on the psychological reality of image schemas (Gibbs & Colston, 2006[1995]), as well as the argumentation presented in Antović (2013), the present research will stress the importance of the generic space constructed at the level of the image schematic structure which serves as a conceptual primitive, and in that sense underlies much of our everyday routine, including the construction and understanding of metaphorical language. With this in mind, the underlying image schematic structure that makes up the generic space is meant to reflect the fact that the notions of categorization and the active processes of on-line reasoning and meaning construction are in fact grounded in the abstract preconceptual knowledge structures with a marked experiential basis, which make up the bulk of our background image schematic knowledge. This suggests that these background knowledge structures at the image schematic level are also available for recruitment, like frames or cognitive models, which are ready to be activated. However, while the entrenched frames are often products of previous conceptual integration, i.e. they are often pre-built compressions ready for recruitment (in the sense of Fauconnier & Turner, 2008, and Fauconnier, 2008), the image schematic background knowledge structures are actually uncompressed conceptual primitives built on experiential interaction during the pre-linguistic phase of development, which can be further enriched via metaphorical projections.

In that sense, the image schematic structure of the generic space can help explain the alignment of the topologies of the input spaces of conceptual integration networks and it serves to shed more light on the process of conceptual integration itself. Therefore, although an artifact of analysis, i.e. a theoretical analytic tool, the generic space can offer an additional insight into the process of conceptual integration. Additionally, the image schematic structure of this “artifact” supplies it with some actual psychological plausibility. Furthermore, in case of the analysis of conventional, i.e. entrenched conceptual metaphors and their corresponding metaphorical expressions, which the present paper is concerned with, such generic space structure can serve to both emphasize, and further explain the nature and structure of such entrenched cognitive mechanisms.


2.2.3. Vital Relations

Vital relations stand for salient conceptual relations that occur between the inputs in a conceptual integration network (Fauconnier & Turner, 2002), which also serve to define the topology of those inputs (Fauconnier & Turner, 2000). Common vital relations include the following: Change, Identity, Time, Space, Cause-Effect, Part-Whole, Representation, Role, Analogy, Disanalogy, Property, Similarity, Category, Intentionality, and Uniqueness (Fauconnier & Turner, 2002: 101; Turner, 2007: 381).

Change is perhaps best illustrated in the “dino-blend” discussed in Fauconnier and Turner (2002: 93-95). Namely, this blend pertains to the evolutionary journey from dinosaurs to birds. Along this path, there are various dinosaurs at different evolutionary stages of development, which, in this particular case, embody the vital relation of Change. In the blend, Change is compressed into Uniqueness, which enables us to grasp the full extent of the evolutionary transformation.

Identity “is not only a vital relation, but perhaps the primary vital relation without which the others are meaningless” (Fauconnier & Turner, 2000: 299; Fauconnier & Turner, 2002: 115). Furthermore, “human mental life is unthinkable without continual compression and decompression of Identity” (Fauconnier & Turner, 2000: 299), and these compressions and decompressions of Identity are also argued to facilitate the functionality of the linguistic system (Fauconnier & Turner, 2000). Although usually understood as a primitive, the vital relation of Identity is in fact “a feat of the imagination, something the imagination must build or disassemble” (Fauconnier & Turner, 2002: 95). For instance, as discussed in Fauconnier and Turner (2002: 96), Identity can serve to connect roles (e.g. the roles of president) in different inputs, which allows us to construe utterances like: “In France the president is elected for a term of seven years, whereas in the US he is elected for a term of four years.” As a result, “the identity connection in this case 11 is manufactured” (Fauconnier & Turner, 2002: 96), since the roles of president are different in the two countries. Additionally, the vital relation of Identity can also be established between more abstract causal structures of the inputs (Fauconnier & Turner, 2002).

Another example of the vital relation of Identity and its compression into Uniqueness is provided by the Ghosts of Predator’s Past example (Fauconnier & Turner, 2000: 299-302; Fauconnier & Turner, 2002: 115-119). Namely, this example describes the American pronghorn which manages to outrun all of the predators over the course of evolution. The conceptual integration network in question has one input space with a modern pronghorn that runs fast, and another input with a prehistoric pronghorn that runs fast to escape predators, where each of the pronghorns is “an identity compression over an epoch in the history of the species” (Fauconnier & Turner, 2000: 300). This identity compression serves to turn “a group identity into a particular identity” (ibid.: 300), since each of the two pronghorns is ‘picked’ from one of the two respective groups of prototypical pronghorns, i.e. the group of pronghorns from the prehistoric epoch, and the group of

11 My italics.


pronghorns from the modern epoch. In the blend, the vital relations of Identity and Intentionality between the pronghorns from the two inputs are compressed into Uniqueness and inner-space Intentionality (Fauconnier & Turner, 2002).

Time relates to “memory, change, continuity, simultaneity and nonsimultaneity, as well as to our understanding of causation” (Fauconnier & Turner, 2002: 96). As seen in the Regatta (Fauconnier & Turner, 2006[1998]: 327-329; Fauconnier & Turner, 2002: 63-65) and the Debate with Kant (Fauconnier & Turner, 2006[1998]: 315-317; Fauconnier & Turner, 2002: 59-62) blends, different times from the input spaces are projected into the same moment in time in the blend. These two examples can also be used to exemplify the vital relation of Space, and its compression. Namely, different physical spaces from the inputs are brought together as a single space in the blend.

Fauconnier and Turner (2002: 96) discuss the example of a fire in a fireplace, and argue that the Cause-Effect vital relation requires that two mental spaces be set up. One space contains the burning logs, and the other the cold ashes. In addition to the Cause-Effect, the inputs are also connected by Time, Space, and Change.

The Part-Whole vital relation may be exemplified by the police trying to identify a murder victim based on a set of fingerprints. One input contains the entire person, while the other input contains the person’s fingerprints, as the most salient characteristic. In the blend, the prints and the person are fused, solving the problem of the person’s identity. This way the outer-space vital relation of Part-Whole is compressed into Uniqueness in the blend.

The vital relation of Representation is closely linked to the trigger-target relation of the mental space theory. For example, one of the inputs can contain a photograph of a person, which can serve as a representation, while the other input can contain the actual person being represented. In the blend, the two are compressed into Uniqueness, allowing us to refer to the representation in the photograph as the actual person.

Role “is a ubiquitous vital relation” (Fauconnier & Turner, 2002: 98), and it can serve to connect roles from one input with values from another input, like in the simplex network Paul is the father of Sally (Fauconnier & Turner, 2002: 120-122). Some other roles can include, for instance, president, pope, or emperor, and they are filled by appropriate values depending on the context.

The vital relation of Analogy is established when “through blending, the different blended spaces have acquired frame structure in common” (Fauconnier & Turner, 2002: 99). Additionally, the vital relation of Analogy is conditioned by the previously discussed Role-Value compression. Analogy can also be compressed, and these are usually compressions into Uniqueness and Change (ibid.). Fauconnier and Turner (2002: 99) discuss the example: Stanford is a West Coast “analogue” to Harvard. Both networks have the frame of a university and the role of a prestigious university. As the result of conceptual integration, one blend acquires the role of Stanford from its other input, and the other blend that of Harvard. Consequently, the two blends are linked via Analogy.


Disanalogy is both grounded on, and coupled with Analogy (Fauconnier & Turner, 2002). The values from the inputs, which are disanalogous, can be introduced into the blend as separate when there is a need to fill multiple slots that have the same role. For example, this is the case in the Regatta blend, when the disanalogous boats are both brought into the blend.

The vital relation of Property is used to assign a certain property to an object. For example, a white shirt has a property white, a convicted felon has a property guilty, etc. Additionally, “blending often compresses an outer-space vital relation ... into an inner-space vital relation of Property in the blend” (Fauconnier & Turner, 2002: 100).

Similarity “is an inner-space Vital Relation linking elements with shared properties” (ibid.: 100). For example, two pieces of fabric can be dubbed similar in terms of their color, or the cross-space relation of Analogy can be compressed into Similarity in the blend.

Category is similar to Property. For example, in the case of a computer virus, the outer-space vital relation of Analogy between a real virus and a computer program is compressed into Category in the blend (Fauconnier & Turner, 2002).

Intentionality includes a group of vital relations linked with “hope, desire, want, fear, belief, memory, and other mental attitudes and dispositions directed at content” (ibid.: 100). Intentionality is very important because it enables us to frame an event as either intentional or unintentional. For example, in the Regatta blend (Fauconnier & Turner, 2006[1998]: 327-329; Fauconnier & Turner, 2002: 63-65), the race frame introduces the relation of Intentionality. Additionally, Intentionality can be introduced into the blend through compression (Fauconnier & Turner, 2002: 100).

As may be transparent from the discussion in the previous paragraphs, the most notable characteristic of the vital relation of Uniqueness is that many vital relations are compressed into it in the blend.

The above listed vital relations are prominent in various types of conceptual integration networks. The present paper will address special attention to the vital relations occurring in metaphorical networks and to their compressions. 2.2.4. Compression

Mental spaces and conceptual integration networks created as discourse and thought unfold are the result of meaning construction, and the need to achieve understanding at a human scale12 (Fauconnier & Turner, 2002). One of the most valuable assets in this process is compression which represents “a phenomenon in conceptual integration that allows human beings to simultaneously control long diffuse chains of logical reasoning and to grasp the global meanings of such chains” (Fauconnier & Turner, 2000: 283). In other words, “compression is used to describe an entity in a blended space that has distinct counterparts in

12 For more details concerning the notion of human scale consult subsection 2.2.5.


multiple input spaces, and, moreover, those counterparts are related to one another via a vital relation” (Coulson & Oakley, 2005: 1532-1533).

The importance of compression as one of the central phenomena of blending was initially discussed in Fauconnier and Turner (2000; 2002), and later further elaborated in Fauconnier (2008) and Fauconnier and Turner (2008). As discussed in Fauconnier and Turner (2002: 312-322), compression can involve scaling or syncopation of a single vital relation. This can include temporal compressions, scaling of cause and effect, etc. Additionally, one of the most important traits of the process of compression is the compression of one vital

relation into another. For example, outer-space vital relations of Representation and Part-Whole can be compressed into Uniqueness in the blend, which is evident when a photograph is used to represent a person. Additionally, as a result of conceptual integration, an outer-space vital relation between the inputs can be compressed either into an inner-space vital relation of the same kind, or of a different kind in the blend (Fauconnier & Turner, 2000).

Compression can also be used to create new vital relations. Namely, conceptual integration can result in new vital relations that did not hold in the inputs, but are present in the blend, and which are obviously closely linked to the emergent structure. If we consider the Regatta blend (Fauconnier & Turner, 2006[1998]: 327-329; Fauconnier & Turner, 2002: 63-65), the two inputs contain the two boats on their journeys, while the blend recruits an additional frame of race, thus introducing the vital relation of Intentionality (in this case the intention to win the race), not present in either of the inputs. A very important form of compression pertains to borrowed compressions, where one of the input spaces already has its own compression which is further exploited in the new blend. This is especially interesting in conceptual metaphors, where, as argued in Fauconnier and Turner (2008), and Fauconnier (2008), inputs in metaphorical networks are often results of previous conceptual integration, i.e. inputs themselves can be instances of entrenched blends.

Compression in conceptual integration networks is generally achieved in two ways: 1) through syncopation; and 2) through scaling. Syncopation means “to leave out significant chunks in a vital relation” (Fauconnier & Turner, 2000: 298), and it can apply either to elements between different inputs, or to elements of the same input (Fauconnier & Turner, 2000). As discussed in Fauconnier and Turner (2000: 298), if we try to think of a good response to criticism that we had received a couple of years earlier, we are actually syncopating over time, and focusing only on the time the criticism was received and the present moment, leaving out everything that happened in between. Unlike syncopation, scaling “preserves the topology of what is compressed” (Fauconnier & Turner, 2000: 298), and the interval that will be scaled can be either inner-space, or outer-space. For example, in the Baby’s Ascent example (Fauconnier and Turner, 2000: 294-297), the entire lifetime is scaled down to the period required to climb the stairs, which occurs in one mental space. In the Bypass blend (Fauconnier & Turner, 2000: 288-292; Fauconnier & Turner, 2002: 65-67), on the other hand, the time interval is scaled down over the two input spaces. Despite their differences, “scaling and syncopation often work together” (Fauconnier & Turner, 2002: 314). For instance, in the dino-blend, there is


the scaling of Change and Time, but there is also syncopation since only the key moments on the evolutionary path are depicted (Fauconnier & Turner, 2002).

One of the main goals in conceptual integration networks is to achieve human scale, 13 which is sanctioned by compression, and the fact that some vital relations are scalable, while those that are not scalable can be compressed to scalable relations (ibid.). Another important trait of compression includes highlights compression, which reflects the requirement that blends and generic spaces need to “reflect the structure and highlights of the overarching stories” (ibid.: 320). Highlights compression can include “compression to Category, compression to Property, and syncopation over detail” (ibid.: 325).

As argued in Fauconnier and Turner (2002: 324-325), and Turner (2007: 382), the above presented features of compression can be summarized under the Governing Principles for Compression which include: 1) borrowing for compression; 2) single-relation compression by scaling; 3) single-relation compression through syncopation; 4) compression of one vital relation into another; 5) scalability; 6) creation by compression (creating a new vital relation in the blend); and 7) highlights compressions.

2.2.5. Optimality Principles14

Optimality principles “characterize strategies for optimizing emergent structure” (Fauconnier & Turner, 2002: 311), and, paired with the constitutive principles, place strong constraints on the process of conceptual integration. The constitutive principles, which refer to the structure and dynamics of conceptual integration, include “partial cross-space mappings, selective projection to the blend, and the development of emergent structure in the blend” (ibid.: 310). The main goal that both constitutive and optimality principles aim to achieve is:

Achieve Human Scale, with its subgoals:

Compress what is diffuse. Obtain global insight. Strengthen vital relations. Come up with a story. Go from many to one. (Fauconnier & Turner, 2002: 322-323; Turner, 2007: 383)

Human scale refers to “the level at which it is natural for us to have the impression that we have direct, reliable, and comprehensive understanding” (Fauconnier & Turner, 2002: 323), and it is such understanding that creates the feeling of global insight. In brief, the entire process of conceptual integration is fueled by the necessity of understanding the world at a level that is both tangible and intelligible. Additionally, compression and human scale go hand in hand, since “compression is a way to achieve human scale” (ibid.: 323), and achieving human scale requires compression. The process of conceptual integration will also lead to the 13 For more details concerning the notion of human scale consult the subsection 2.2.5. 14 The present section will provide only a basic overview of Optimality Principles. For a more detailed account, see Fauconnier & Turner, 2002: 327-333, and Turner, 2007: 382).


strengthening of vital relations “either by making new ones, or by intensifying existing ones, or by converting vital relations of one type into vital relations of another type” (ibid.: 323). The blend can also provide a simple

story for the entire network, like in the Regatta15 example where the blend supplies the simple story of a race (Fauconnier & Turner, 2002). Fauconnier and Turner (2002: 287-288) also discuss the example of the Story of

Life, where the Story of Life serves as a preexisting overarching integration that guides some of the compressions in the Birth Stork Network,16 i.e. it serves to provide a story that facilitates the development of the conceptual integration network. Consequently, each of the blends in the Birth Stork Network has a structure that is aligned with the preexistent compression of the Story of Life. Namely, in each of the blended spaces, “states are locations, and there is a change from one ranked state/location to another” (Fauconnier & Turner, 2002: 289). Finally, as a direct result of compression, conceptual integration networks are expected to go from many elements in the input spaces to a fewer number of elements in the final blend (Fauconnier & Turner, 2002).

Optimality principles include the following: ■ The Topology Principle: Other things being equal, set up the blend and the inputs so that useful topology in the inputs and

their outer-space relations is reflected by inner-space relations in the blend. (Fauconnier & Turner, 2002: 327; Turner, 2007: 382)

■ The Pattern Completion Principle: Other things being equal, complete elements in the blend by using existing integrated patterns as additional inputs. Other things being equal, use a completing frame that has relations that can be the compressed versions of the important outer-space vital relations between the inputs. (Fauconnier & Turner, 2002: 328; Turner, 2007: 382)

■ The Integration Principle: Achieve an integrated blend. (Fauconnier & Turner, 2002: 328; Turner, 2007: 382)

■ The Maximization of Vital Relations Principle: Other things being equal, maximize vital relations in the network. In particular, maximize the vital relation in the blended space and reflect them in outer-space vital relations. (Fauconnier & Turner, 2002: 330; Turner, 2007: 382)

■ The Intensification of Vital Relations Principle: Other things being equal, intensify vital relations. (Fauconnier & Turner, 2002: 330; Turner, 2007: 382)

■ The Web Principle: Other things being equal, manipulating the blend as a unit must maintain the web of appropriate connections to the input spaces easily and without additional surveillance or computation. (Fauconnier & Turner, 2002: 331; Turner, 2007: 382)

■ The Unpacking Principle: Other things being equal, the blend all by itself should prompt for the reconstruction of the entire network. (Fauconnier & Turner, 2002: 332; Turner, 2007: 382)

■ The Relevance Principle: Other things being equal, an element in the blend should have relevance, including relevance for establishing links to other spaces and for running the blend. Conversely, an outer-space relation between the inputs that is important for the purposes of the network should have a corresponding compression in the blend. (Fauconnier & Turner, 2002: 333; Turner, 2007: 382)

■ The Compression Principle:17 Achieve compressed blended spaces. (Turner, 2007: 382)

Grady, Oakley, and Coulson (2007[1999]) discuss an additional principle of Metonymic Tightening, according to which “relationships between elements from the same input should become as close as possible within the blend” (ibid.: 426).

In sum, optimality principles have much to do with vital relations and their projections, and by offering additional constraints to those of the constitutive principles, optimality principles are able to dictate the course

15 For details see Fauconnier and Turner, 1994; 2006[1998]; 2002. 16 For details see Fauconnier and Turner, 2002: 284-289. 17 The governing principles for compression have already been presented in section 2.2.3. above.


of the development of conceptual integration networks. Additionally, when it comes to various types of networks, many of these principles interact and compete, so that the satisfaction of one of the principles may imply a violation of other principles. 2.2.6. Basic Types of Conceptual Integration Networks18

The four basic types of conceptual integration networks include: simplex, mirror, single-scope, and double-scope networks19 (Fauconnier & Turner, 2006[1998]; 2002). A notion closely related to the types of integration networks that also serves as the main criterion for the above given classification is that of an organizing frame. Namely, an organizing frame of a mental space is “a frame that specifies the nature of the relevant activity, events, and participants” (Fauconnier & Turner, 2006[1998]: 341; Fauconnier & Turner, 2002: 104), or in other words, “an organizing frame provides a topology for the space it organizes by providing20 a set of organizing relations among the elements in the space” (Fauconnier & Turner, 2006[1998]: 341; Fauconnier & Turner, 2002: 123). Although the notion of frames has already been introduced in section 2.1.7, it is worth noting at this point that in terms of the Mental Space Theory, frames are understood as “entrenched mental spaces that we can activate all at once” (Fauconnier, 2007: 352; Fauconnier & Turner, 2002: 103). Furthermore, even entire networks of spaces with their blends can become entrenched and available for recruitment (Turner, 2007; Fauconnier & Turner, 2002). 2.2.6.3. Single-scope Networks

The main characteristic of single-scope networks is that inputs have different organizing frames, and only one of those frames is projected as the organizing frame of the blend (Fauconnier & Turner, 2006[1998]; 2002; Coulson, 2001). As a result, “the projection to the blend in a simple single-scope network is highly asymmetric” (Fauconnier & Turner, 2002: 127). Also, with their structure in mind, single-scope networks are prototypically metaphorical since they “give us the feeling that one thing is giving us insight into another thing” (ibid.: 129), which is reminiscent of Lakoff and Johnson’s (2003[1980b]) definition of metaphor. Fauconnier and Turner (2002: 127-129) propose two types of single-scope networks: 1) the inputs are not contained within a larger history; and 2) the inputs are contained within a larger history. In the former type of single-scope networks, there are no Identity connections between the inputs, nor are there vital relations of Time, Space, Change, Cause-Effect, and Intentionality that license direct connections between the inputs. In short, there is no obvious relevance of one input space for the other. In the latter type of single-scope networks, the inputs 18 This section will describe only single- and double-scope networks, since those are typical instances of metaphorical networks with which the present research is concerned. For a detailed account of simplex and mirror networks see Fauconnier & Turner, 2002. 19 Note that Seana Coulson (2001: 120-122) uses different terminology to describe these same types of networks. Namely, she discusses single framing networks, frame networks, one-sided networks, and two-sided networks, which correspond to simplex, mirror, single-scope, and double-scope networks, respectively. 20 My italics.


are relevant to each other, and there are outer-space vital relations that connect them. The connections can be established between the respective organizing frames, as well as between elements below the frame level.

Considering that inputs have different organizing frames, conceptual clashes in single-scope networks are extremely transparent, and the blend normally inherits the organizing frame of the framing input, i.e. the source input. In this type of networks the clash is resolved by projecting only one of the organizing frames to the blend. Typically, the framing input “has a prebuilt superb compression that is exploited to induce a compression for the focus input” (ibid.: 129), and the use of such preexistent compressions is in general a salient property of single-scope networks. In other words, single-scope networks satisfy the principles of borrowing for compression and scaling compression. However, “the topologies of the two spaces are preserved in the cross-space mapping” (ibid.: 130), in spite of the clashes.

Figure 3. Boxing CEOs (adopted from Fauconnier and Turner, 2002: 128)

In terms of optimality principles, single-scope networks satisfy integration owing to the “borrowed”

compression from the source input, which in turn satisfies the topology principle. The topology principle is further satisfied in projections from the target input because the topologies of the source and target inputs are already ‘prealinged’ by the presence of a conventional metaphor (Fauconnier & Turner, 2002). The metaphorical nature of the mappings also satisfies web, while unpacking goes on in a similar fashion as in the case of mirror networks.21 Namely, unpacking is facilitated by the fact that the blend is not integrated to the same extent at the more specific levels as it is at the more general level.

Let us consider the example of Boxing CEOs, presented in Fauconnier and Turner (2006[1998]: 344; 2002: 126-131). This metaphor of CEOs as boxers allows us to compress our understanding of business

21 For more details on mirror networks, consult Fauconnier and Turner, 2002.


competition. In that sense, cross-space connections can be established between the CEOs and boxers, punches and business moves, and the boxing match and business competition. This is also an instance of a network in which the inputs are not contained within a larger history, since there is no direct obvious link between them. The conceptual clash is resolved by projecting the organizing frame from the source input22 (boxing) to the blend, which in turn introduces the preexistent compression of boxing into the blend. This way, structure from the target input (business) is “projected into the already compressed inner-space relations ... from the framing input” (Fauconnier & Turner, 2002: 129). Additionally, “cause-effect relations, agent-action relations, and temporal ordering are aligned in the two spaces” (ibid.: 130), owing to which the projection of elements from the target input as values into the source organizing frame does not disrupt the structure of the organizing frame. In short, the blend achieves the conceptual integration of the events from the target input into a unit. Additionally, in terms of compressions, it is worth noting that the outer-space vital relation of Analogy is compressed into an inner space vital relation of Uniqueness: in the blend, CEOs are fused with boxers. 2.2.6.4. Double-scope Networks

Double-scope networks constitute fertile ground for conceptual clashes, owing to the fact that their inputs have different organizing frames, and both of those frames contribute to the organizing frame of the blend (Fauconnier & Turner, 2006[1998]; 2002; Coulson, 2001). As discussed in Fauconnier and Turner (2002: 340-345), since the blend has a specially developed organizing frame of its own, Compression, Topology, Integration, and Web principle are not automatically satisfied in double-scope networks. This increases the possibilities for competition between the optimality principles and it becomes more difficult to satisfy them. As a result, the construction of a functional double-scope blend requires that some optimality principles be relaxed. The fact that all the spaces in the network have different organizing frames facilitates the Unpacking Principle. Namely, in order for the elements from the blend to be projected back to the inputs, first the organizing frame of the blend needs to be unpacked into the organizing frames of the inputs.

Double-scope networks also include networks with high asymmetry and nonclashing double-scope

networks.23 In the former type of networks, although the organizing frame of the blend adopts elements from the organizing frames of both inputs, it is basically an extension of the organizing frame of only one of the inputs. In the latter type of networks there is no clash between the organizing frames of the inputs, and the blend contains both of them.

Among other examples of double-scope networks, Coulson (2001: 168-172) discusses the idiom Digging Your Own Grave, also described in Fauconnier and Turner (2006[1998]: 321-323; 2002: 131-135). The clash of topologies between the two inputs of this network is manifested on multiple levels, including

22 Note that similar to the CMT framework, inputs in metaphorical integration networks are also labeled source and target. 23 For a detailed account of these two special types of double-scope networks see Fauconnier & Turner, 2002: 134-135.


“causality, intentionality, participant roles, temporal sequence, identity, and internal event structure” (Fauconnier & Turner, 2002: 132). The concrete structure of the blend is adopted from the “grave” input, whereas the event structure, intentionality, and causality come from the “failure” input. The clash in this network is presented in terms of the shift in the causal structure owing to which the existence of a grave causes death, and the other way around. This causal structure is adopted from the “failure” input, while death and dying come from the “grave” input. As a result, the amount of trouble maps onto the depth of grave, and making bad decisions corresponds to digging the grave. Such framing also fuses the roles of agent and patient, and the patient who is also the agent digs his own grave. The resulting emergent structure also has backward projections to the “failure” input, where it creates real-life inferences.

I propose another real-life example of double-scope blending, based on my personal experience. Namely, imagine a situation where a man is in a long-term relationship with a woman seven years younger than he is. Now imagine that the woman has just graduated from the University and the couple is out celebrating. At one point, the man turns to the woman and says: “Now I’m your Mrs. Robinson.” To fully appreciate this statement, one should be familiar with the movie The Graduate (1967), starring Dustin Hoffman and Anne Bancroft.

Figure 4. Mrs. Robinson

The conceptual integration network that describes this example has two inputs, the movie space and

the reality space, while the generic space contains elements that are shared by the entire network (see Figure 4 for details). Elements Mrs. Robinson and the man, the graduate and the woman, and their age differences are connected by the outer-space vital relations of Analogy. The extramarital affair and the long-term


relationship are on the other hand linked by the vital relation of Disanalogy. As the result of compression, the blend develops its central emergent structure. Namely, the outer-space vital relations of Analogy between the man and Mrs. Robinson, and the woman and the graduate are projected into Uniqueness in the blend. However, the two compressed elements in the blend acquire their genders and specific values (i.e. their real identities24) from the reality space, whereas their roles are acquired from the movie space. In that sense, the blend contains Mrs. Robinson, who is in fact a man, and the graduate, who is in fact a woman. Furthermore, the age difference and the long-term relationship status are projected directly from the reality space. Additionally, Time from the movie and Time from reality are also compressed through syncopation into the same moment in time in the blend. In keeping with the partial nature of projections, the languages used in the two respective input spaces are irrelevant, and not projected from either of the inputs (like in the Debate with

Kant example discussed in Fauconnier & Turner, 2002). The arguments presented in the previous paragraph lend support to the conclusion that the resulting

topology of the organizing frame of the blended space is in fact a direct result of interaction between the organizing frames of the two input spaces, thereby rendering the present network double-scope. Namely, while the two compressed elements assume their roles from the ‘movie space’, their other characteristics are acquired from the ‘reality space’. Furthermore, the age difference and the relationship status are also acquired from the ‘reality space’, thereby licensing a novel, emergent structure that appears in the form of a long-term

relationship between Mrs. Robinson and the graduate. Prior to the process of conceptual integration, this structure was not available in either of the two inputs.

2.2.7. CBT and Metaphor

In the domain of the CBT, metaphors normally appear as single- or double-scope networks. Unlike the unidirectional source-target mappings stipulated by the CMT, the CBT introduces at least the minimal network model, allowing projections in multiple directions, including backward projections from the blend to the inputs. In their paper Rethinking Metaphor, Fauconnier and Turner (2008) argued for a revised approach to metaphor analysis. The authors suggested that a full account of conceptual metaphors needs to include a greater number of elements than a simple four-space-model. Namely, the inputs of metaphorical networks are rarely simple constructs, and they usually represent previously constructed entrenched blends available for recruitment (i.e. borrowed compressions). This means that a seemingly simple conceptual integration network can in fact be rather complex and elaborate. In other words, “analysis of metaphor requires analysis of elaborate integration networks producing what can seem like straightforward mappings between two domains taken as primitives” (Fauconnier & Turner, 2008: 61).

24 Note that “identity” here is not being used to label the vital relation of Identity.


Building on the analysis of the metaphor TIME IS SPACE Fauconnier and Turner (2008: 60) conclude that time and time-space mapping appear as the emergent structure in the blend, rather than time being a simple primitive input. Additionally, “the topology of the domain for time is incompatible with the domain of objects moving in space” 25 (Fauconnier & Turner, 2008: 56), owing to which meaning construction in metaphorical expressions that correspond to the above introduced conceptual metaphor is ultimately a result of resolving the clashes that appear due to the inconsistencies in the topologies of the two inputs. Another important notion that needs to be stressed includes the differences between the universal event and subjective experience. Namely, the universal event assumes the same order of events for all participants, with time as an invariant component. Depending on the specific context, however, different participants can construe the duration of a single event as being different, which gives way to expressions such as: “These three hours went

by slowly for me, but the same three hours went by quickly for him” (Fauconnier & Turner, 2008: 61). Meaning construction in cases like this one is “a result of resolving a clash between subjective experience and shared universal events” (Fauconnier & Turner, 2008: 64).

In sum, Fauconnier and Turner (2008: 54) propose that the following notions be taken into consideration during metaphor analysis: ■ Integration networks: Blends always involve multiple mappings. ■ Cobbling and sculpting: Integration networks always have parts that are entrenched (i.e. conventional),

and parts that involve novel mappings and compressions. As a result, integration networks that permeate everyday life are usually a mix of these two components.

■ Compression: “Integration networks achieve systematic compressions” (Fauconnier & Turner, 2008: 54). Although compression and the notion of borrowed compressions has already been addressed in their previous work (e.g. Fauconnier & Turner, 2000; 2002), the authors here argue that the systematic nature of compression has been overlooked in previous research. Such importance of compression is also stressed in Fauconnier’s 2008 lecture How Compression Gives Rise to Metaphor and Metonymy. In short, the systematic nature of compression is reflected in the fact that “the ability to use standard techniques and patterns of compression and decompression enables us to work at once over elaborate integration networks” (Fauconnier & Turner, 2008: 54).

■ Inference: When metaphors involve multiple inputs, the topologies of those inputs will often clash. Consequently, some elements from those inputs will not be projected to the blend, which means that “inference transfer is not in itself the driving force behind metaphor” (ibid.: 54).

■ Emergent structure: The authors argue that some of the previous approaches to metaphor analysis neglected to recognize the fact that many of the most fundamental metaphorical projections (e.g. the above discussed TIME IS SPACE), are in fact emergent.

25 For a detailed list of the points of incompatibility between the two inputs see Fauconnier and Turner, 2008: 56-57.


Grady, Oakley, and Coulson (2007[1999]: 426-427) discuss the SHIP OF STATE metaphor, and they suggest that although this metaphor represents an instance of a conventional mapping, it is also related to more fundamental conceptualizations, like ACTION IS SELF-PROPELLED MOTION, TIME IS MOTION, etc. This suggests that simple metaphors can be joined to pave the way for more elaborate mappings. The authors also argue that “one blend may be the input for another” (Grady, Oakley, & Coulson, 2007[1999]: 427), which reflects Fauconnier and Turner’s notion of the systematic nature of compression. Additionally, Grady, Oakley, and Coulson (2007[1999]) argue one of the basic characteristics of metaphorical networks to be fusion with accommodation, which means that not all elements from the inputs are fused in the blend. Essentially, in the case of metaphor, fusion corresponds to the compression of outer-space vital relations of Analogy into inner-space vital relations of Uniqueness in the blend. This way, “a single element in the blend corresponds to an element in each of the input spaces” (Grady, Oakley, & Coulson, 2007[1999]: 431). 2.2.8. CMT vs. CBT: Towards a Unitary Framework

In view of the CMT, “a conceptual metaphor consists of a (partial) mapping of the basic structure of one conceptual domain (the source) onto another (the target)” (Turner & Fauconnier, 1995: 184). CBT, on the other hand, introduces a four-space model, in which projections are not restricted to unidirectional source-target mappings advocated by the CMT. Instead, elements from both the source and target inputs are projected to the blend, with an additional possibility of backward projections that can affect inferences in the input spaces. Another important difference between the two frameworks is presented in the fact that the domains of the CMT framework pose as stable, entrenched structures, while mental spaces from the CBT framework appear as dynamic constructs, structured by frames, domains, and cognitive models. Additionally, the CBT framework is usually aimed at investigating instances of novel conceptualizations, whereas the CMT deals with entrenched structures (Grady Oakley, & Coulson, 2007[1999]; Coulson, 2006b).

Despite the obvious incoherencies between the two frameworks, there are also certain similarities. Namely, metaphor is treated as a conceptual phenomenon in both frameworks; both frameworks deal with systematic projections between conceptual domains; both frameworks introduce certain constraints; etc. (Grady Oakley, & Coulson, 2007[1999]). As discussed in Fauconnier and Turner (1994: 4-5), and Turner and Fauconnier (1995: 184), the two space model of the CMT is in fact only a border-line case of the more elaborate many-space model of the CBT that “explains a range of phenomena invisible or untreatable under the two-domain model and reveals previously unrecognized aspects of even the most familiar basic metaphors” (Turner & Fauconnier, 1995: 184). Furthermore, Grady, Oakley, and Coulson (2007[1999]) in their discussion on Blending and Metaphor also suggest that the two approaches should be treated as complementary, and argue that “the conventional conceptual pairings and one-way mappings studied within CMT are inputs to and constraints on the kinds of dynamic conceptual networks posited within BT” (Grady,


Oakley, & Coulson 2007[1999]: 436). Similar positions are adopted in Coulson (2006a; 2006b). Such treatment of the CMT as a special case of the CBT, instead of being dubbed a separate theoretical framework, allows the CMT to still serve as a useful aid in the research in conceptual metaphors.

Grady, Oakley, and Coulson (2007[1999]: 422-424) discuss the well-known example of the SURGEON

IS A BUTCHER metaphor (Figure 5), and the analysis of this metaphor reveals the advantages of the many-

space model of the CBT. Namely, the blend has the emergent structure of incompetence that is difficult to explain relying merely on the CMT model. The problem lies in the fact that both surgeon and butcher are highly competent professionals. Although butchery might be dubbed a less prestigious calling compared to surgery, there are no obvious elements that could serve to discredit the competence of the former. Still, the above presented metaphor is perfectly clear, and the inference of the surgeon’s incompetence is retrieved without difficulties.

Figure 5. Surgeon is a butcher (adopted from Grady, Oakley, and Coulson, 2007[1999]: 423)


The CBT framework provides a plausible account of the emergent structure. Namely, the role of butcher is projected from the source input, while elements like the identity of the patient and identity of the surgeon are projected from the target input. The clash between the goals of a surgeon and means and manner of a butcher results in the emergent structure in the blend that is not available in either of the inputs. In short, “the incongruity of the butcher’s means with the surgeon’s ends leads to the central inference that the butcher is incompetent” (Grady Oakley, & Coulson, 2007[1999]: 424). In that sense, while the CMT allows the identification of the source and target inputs, and establishes connections between their elements, the CBT expands that relationship by analyzing it in the four-space network of conceptual integration that also includes the emergent structure not captured by the former framework. In sum, while the CMT model is focused on investigating entrenched conceptualizations, the CBT model deals with dynamic on-line processes of meaning construction.

With the arguments presented in the above paragraphs in mind, it can be concluded that the joint use of the two frameworks can facilitate and unite the investigation of both conventionalized mappings available for recruitment, and individual instances of active, on-line meaning construction. 2.2.9. Taking a Slightly Different Turn: The Coded Meaning Model

In addition to the approach to conceptual blending advocated by Fauconnier and Turner, the present paper will discuss another ‘alternative’ approach - The Coded Meaning Model, introduced by Coulson and Oakley (2005). Namely, building on the previous work in conceptual blending performed by Fauconnier and Turner (e.g. 2002), and Brandt and Brandt (2002), 26 Coulson and Oakley (2005) introduced the coded

meaning model. The predominant role of this approach was to attempt to foreground the importance of contextual information and background cognition that play vital roles in the process of meaning construction. Additionally, the authors argue that “coded meaning plays an important role in the construction of conceptual integration networks” (Coulson & Oakley, 2005: 1511).

The coded meaning model includes a presentation space, a reference space, a blended space, and a grounding box. The presentation space is similar to the source domain of a conceptual metaphor, while the reference space “represents the facet of a situation that is the present focus of attention of the discourse participants” (ibid.: 1516). The grounding box is used to include the “role of implicit and explicit assumptions in meaning construction” (ibid.: 1517). In other words, following in the footsteps of the mental space theory, the role of the grounding box is to account for the influence of the immediate context, as well as for the background knowledge that might be recruited for the purposes of building a conceptual integration network. 26 Brandt and Brandt (2002) developed a model that consists of six spaces: a semiotic space, a presentation space, a reference space, a relevance space, a virtual space, and a meaning space. Such approach was developed as a result of criticism of the blending theory, specifically its inability to provide a clear account of the origin of the emergent structure (cited in Coulson & Oakley, 2005: 1515). Coulson and Oakley adopted the presentation and reference spaces form their model, while the blended space comes from Fauconnier and Turner’s approach.


As such, the grounding box appears as a ‘partial’ counterpart of the generic space. Namely, while the generic space contains general relations at the conceptual level, which apply to the entire network, the grounding box on the other hand is more concerned with the elements of the immediate context and background knowledge structures that usually take a more tangible form.

Coulson and Oakley (2005: 1517) propose two possibilities for a grounding box: deictic and displaced. The deictic grounding box includes specific conditions like time, place, status of discourse participants, or the forum where the situation is taking place (e.g. an office). A displaced grounding box is set up when the immediate context activates “structured background knowledge that constitutes and constrains the interpretive process” (ibid.: 1517). This is reminiscent of frame-recruitment, where a particular frame introduces not only roles, but also full sets of attitudes, available from long-term memory. As a result, the construed meaning will depend on “the background knowledge of the participants, as well as the extant mental space configuration set up by the local context” (ibid.: 1522).

Figure 6. On the one hand ... on the other hand (adopted from Coulson and Oakley, 2005: 1524) The idiom “on the one hand/on the other hand” discussed in Coulson and Oakley (2005: 1523-1524)

can be used to exemplify an analysis in the above described theoretical model (see Figure 6). The authors discuss a paragraph dealing with a relationship between the EU and Israel that appeared in a newspaper article:

‘‘I noticed people saying, which they didn’t use to say, but now they have been saying for the last year or so, perhaps Israel should have never been invented, but then I hear them say that the only solution is for the Arabs to accept the existence of Israel, and normalize with it’’, she says. ‘‘So I see them on the one hand acknowledging that the Arabs have reason to resent the Europeans for supporting something they find intrusive, but on the other hand demanding the Arabs to put up with it, now


that it has happened. And so it would be more consistent for the Europeans, and the British included, to say: Yes, we did want Israel to exist, and it is not a mistake, but since Israel is more us than you, Israel’s future will be more with us than with you, the Arabs.’’ (Coulson & Oakley, 2005: 1523; originally appeared in Haaertz, January 26, 2004)

The idiom “on the one hand/on the other hand” prompts the construction of a blended space in which each hand supports a certain position, and there is an emergent structure that suggests that the position held in one hand should be dropped in favor of the position in the other hand. This way the blend is actually anchored in the human body (in the sense of Hutchins, 2005), and the two hands provide spatial orientation for the two opposing positions, which is licensed by the entrenched nature of the idiom. The presentation space contains human hands and the bilateral symmetry of the human body, while the reference space contains two ideas: 1) the creation of Israel was a mistake, and 2) Arabs must put up with Israel. Elements of the grounding box, which include participants, forum, and circumstances, are presented in Figure 6.

2.2.10. Section Summary

The present section provided an overview of the basic tenets of the Mental Space Theory, and it also presented the main elements of the Conceptual Blending framework, where blending is viewed as “an offspring of the theory of mental spaces” (Harder, 2005: 1639). Special attention was paid to vital relations, topology principles, and the notions of compression and the systematic nature of compression in the sense of Fauconnier and Turner (2008), and Fauconnier (2008). Additionally, the basic characteristics of single- and double-scope networks which will be of particular interest for the present research have also been presented. The current section also pointed out the main similarities and differences between the CMT and CBT, and presented arguments in favor of a joint use of the two frameworks. Finally, the section finished by presenting the Coded Meaning Model as introduced by Coulson and Oakley (2005).

Affect and Emotions 46

2.3. Affect and Emotions

The present section of the paper will outline some of the dominant approaches to research in the domains of affect and emotional responses. Building on the basic definitions of emotion concepts and affect, the section continues with the discussion of the Russell’s circumplex, the PANAS model, the Core Affect model, and the Conceptual Act Theory of Emotion. This is followed by the overview of literature dealing with the role of emotional and affective experiences in relation to the media and politics. Finally, the section closes by summing up the main conclusions, thereby providing the theoretical grounding that will be used in the second part of the present research. 2.3.1. Emotion Concepts

In general terms, “emotions are commonly defined as positive and negative feelings in certain situations” (Nešić et al., 2009: 137), and the ability to make sense of those feelings is facilitated by the use of concepts. Emotion concepts are similar to other abstract concepts addressed in the context of cognitive semantics, and they “are not only fundamental for an understanding of the social world; they are fundamental to the development of an individual’s behavioral repertoire” (Niedenthal, 2008: 587). In that sense, a complete account of human behavior and experience necessitates a detailed insight into the structure of emotional dimensions of that experience. An important distinction to be stressed at this point is the one between emotions and moods. Namely, although related in the sense that they refer to subjective states, “mood is a much broader concept than emotion” (Gray & Watson, 2007: 171), since moods last considerably longer compared to emotions, and they are also more frequent and consistent. Emotions, on the other hand, “are typically short-lasting, may be intense, focused, and often have a clear causal underpinning” (Barlett & Gentile, 2011: 60).

Among other approaches to the analysis of emotions, Niedenthal (2008) discusses the dimensional models (e.g. Russell, 1980; Watson, Clark, & Tellegen, 1988), and models that view emotion concepts as prototypes (e.g. Russell, 1991; 2003). The dominant dimensional models that discuss the dimensions of valence and arousal, and notions of positive and negative affect, are described in great detail in subsections 2.3.2.1, 2.3.2.2, and 2.3.2.3. The application of prototypicality effects onto emotion concepts reflects the idea that an emotion concept is “a fuzzy one defined by a set of probable features, but not necessary and sufficient ones, that overlap with closely related categories” (Niedenthal, 2008: 592). Further application of the prototype theory to emotion concepts is evident in Russell’s research (1991), where he views emotions as “events having a causal and temporal structure, and not objects, and so the notions of a temporally structured script best captures the representation of an emotion” (Niedenthal, 2008: 593). Consequently, “to know the sense of a term like anger, fear, or jealousy is to know the script for that emotion” (Russell, 1991: 39, cited in Niedenthal,


2008: 593), which reflects the connection between the Russell’s notion of an emotion script, and the notion of frames from cognitive semantics. 2.3.2. Affect

“Affect plays a central role in human experience” (Gray & Watson, 2007: 171) and, in comparison to emotions and moods, it stands as a “more inclusive psychological construct that refers to mental states involving evaluative feelings” (ibid.: 171). Furthermore, “affect is conceptualized as more global than emotions and mood, and may consist of many different emotional and mood states” (Barlett & Gentile, 2011: 60). In view of its value in the construction of the fully developed human scale experience, it can be added that, “perception without an affective component lacks the first-person, subjective quality that is the hallmark of conscious awareness of external sensory information” (Duncan & Barrett, 2007: 1197).

In terms of its structure, the two dominant models of affect include the specific-affect theories, and dimensional theories. The former group of theories was based on a discrete-affect model, where it was considered that a number of well-defined content factors were required to fully explain the affective experience (Gray & Watson, 2007). However, “these specific-affect scales usually were found to be significantly interrelated” (ibid.: 172). The latter group of theories was designed to overcome the shortcomings of the former group, and “given the very broad non-specific nature of affective experience of affect, researchers began to adopt models that bypassed these discrete affects and posited fewer underlying dimensions” (ibid.: 172). Two of the most dominant dimensional models include the Russell’s circumplex, and the PANAS model, both of which are discussed in greater detail in subsections 2.3.2.1 and 2.3.2.2.

Another important fact to be stressed when dealing with the measurements of affect pertains to the difference between state and trait affect. Namely, while state affect refers to short emotional episodes directly related to the immediate context, i.e. a specific stimulus, trait affect reflects “long-term, stable individual differences that reflect a person’s general tendency to experience a particular affective state” (ibid.: 172). In that sense, type-one questionnaires27 used in the second part of the present research were designed to measure state affect.

2.3.2.1. The Circumplex Model of Affect

Russell (1980) introduced the Circumplex Model of Affect, where he proposed that “affective states are, in fact, best represented as a circle in a two-dimensional bipolar space” (Russell, 1980: 1161-1162).28 In that sense, this model represents one of the dimensional models of affect, where affect is represented in terms of two main components: valence (pleasure – displeasure), and arousal (activation - sleep). As a result, “any

27 See section 4.2 for details. 28 See Figures 7 and 8 for details.


affect word could be defined as some combination of pleasure and arousal components” (ibid.: 1163). In short, the circumplex model of affect represents a coordinate system in a two-dimensional space, where the horizontal axis corresponds to the pleasure-displeasure dimension of affect, while the vertical axis corresponds to the arousal-sleep dimension. Consequently, all affect-related words can be defined as vectors that originate from the center of the coordinate system, and are distributed in a circular fashion in relation to that center (Russell, 1980). As a result, each vector defines a single point in the two-dimensional space of the circumplex. This in turn supports Russell’s proposition that there are no clear-cut boundaries between emotion concepts, and they cannot be accurately represented on a simple scale, but require a two-dimensional geometric model. In other words, “each word can thus be considered a label for a fuzzy set, defined as a class without sharp boundaries, in which there is a gradual but specifiable transition from membership to nonmembership” (Russell, 1980: 1165). Such reasoning is also aligned with some of the basic tenets of cognitive semantics, where words are seen as prompts for vast repositories of meaning, rather than exclusive labels that correspond to strictly defined entities (e.g. Saeed, 2003). Additionally, Posner, Russell, and Peterson (2005) provided a more recent account of the application of the circumplex model in which they gave further validation for its use, and, among other points, discussed the neural circuitries that support valence and arousal, as well as the developmental correlates of the circumplex.

Figure 7. Eight affect concepts in a circular

order (adopted from Russell, 1980: 1164)

Figure 8. Direct circular scaling coordinates for 28 affect words

(adopted from Russell, 1980: 1167)

Building on the previously described circumplex model, Russell, Weiss, and Mendelsohn (1989)

introduced the Affect Grid that represents “a scale designed as a quick means of assessing affect along the dimensions of pleasure-displeasure and arousal-sleepiness” (Russell, Weiss, & Mendelsohn, 1989: 493). The Affect Grid is a single-point item scale represented in the form of a 9x9 matrix organized along the valence and arousal dimensions of affect, like the Russell’s circumplex. Affect descriptors listed around the grid “are arranged from the northwest corner of the grid in a clockwise direction – stress, high arousal, excitement,


pleasant feelings, relaxation, sleepiness, depression, and unpleasant feelings – so that the two major dimensions are represented as bipolar opposites”29 (Gray & Watson, 2007: 177). In experimental contexts, the experimental subjects have the task to rate their response to the presented stimulus by marking an appropriate field on the grid. For instance, the affect grid was used as one of the instruments in the study conducted by Lindquist and Barrett (2006).

Figure 9. The Affect Grid

(adopted from Russell, Weiss, and Mendelsohn, 1989: 494) 2.3.2.2. The PANAS Model

The Positive and Negative Affect Schedule (PANAS), first introduced by Watson, Clark, and Tellegen (1988), represents another dimensional model of affect, which distinguishes between the two primary dimensions of mood - positive (PA) and negative affect (NA). To avoid the misinterpretation of equating PA with pleasure and NA with displeasure, it is important to stress the fact that “PA is a measure of the combination of pleasure and arousal, while30 NA is a measure of the combination of displeasure and arousal” (Russell, Weiss, & Mendelsohn, 1989: 498). The measurement tool that this model utilizes is based on a list of adjectives with ten adjectives related to PA and ten adjectives related to NA, where each adjective is graded on a 5-point Likert scale. The final list of adjectives was obtained using factor analysis that was applied to the initial list of 60 terms adopted from previous research. The aim of such a procedure was to “select terms that were relatively pure markers of either PA or NA” (Watson, Clark, & Tellegen, 1988: 1064). Additionally, the PANAS questionnaire can be tailored to suit a particular purpose by introducing an appropriate temporal reference (see Figure 10 for details). Apart from the presently described version of the PANAS model, Watson and Clark (1994) introduced the PANAS-X model, where they “expanded this instrument by including 11 factor-analytically derived scales that assess specific, lower order affects” (Gray & Watson, 2007: 176).31

The fact that “one interesting property of a circumplex is that since rotation of the axes leaves the circular configuration of the variables intact, rotation is considered arbitrary” (Russell, 1980: 1171) offers a connection between the Russell’s circumplex and the PANAS model. Namely, the PA and NA dimensions from

29 See Figure 9 for details. 30 My italics. 31 For more details about the PANAS-X model, consult Watson and Clark, 1994, and Gray and Watson, 2007: 176.


the PANAS model can be obtained directly from the circumplex by rotating the circumplex’s valence-arousal space by 450 counterclockwise (Russell, Weiss, & Mendelsohn, 1989; Gray & Watson, 2007; Niedenthal, 2008). This is made possible owing to the fact that “from a circumplex point of view, any rotation of the axes is possible because the structure of affect is determined by the circular ordering” (Russell, Weiss, & Mendelsohn, 1989: 494). However, it needs to be stressed that these two dimensional models “offer opposing views on the polarity of the dimensions of affective space” (Gray & Watson, 2007: 173). While Russell’s circumplex stresses the bipolar nature of affect, PANAS focuses on two mostly unipolar dimensions.32 Namely, “whether or not PA and NA are in fact orthogonal or correlated dimensions has been, and remains, a controversial issue“ (Thompson, 2007: 229). 33 Additionally, both models have received numerous empirical validations (Gray & Watson, 2007; Lang & Ewoldsen, 2011).

Figure 10. The PANAS

(adopted from Watson, Clark, and Tellegen, 1988: 1070)

Figure 11. I-PANAS-SF

(adopted form Thompson, 2007: 240) 32 For a detailed account of the differences and similarities between these two dimensional models of affect see Gray and Watson, 2007: 172-173. 33 For a discussion on the validity of the PANAS model see, for example, Crawford and Henry, 2004.


In order to aid the advancement of cross-cultural research on affect, building on the initial PANAS model (Watson, Clark, & Tellegen, 1988), Thompson (2007) developed an internationally reliable 10-item version dubbed I-PANAS-SF (see Figure 11 for details). His research “was specifically aimed at developing a measure for trait affect” (Thompson, 2007: 231). An additional motivation for such course of research resides in the fact that covariances between some of the PANAS items “suggested considerable redundancy of the PANAS items closely related to each other in meaning” (ibid.: 230), which gave way for item reduction without compromising scale efficacy and validity. 2.3.2.3. The Core Affect Model

The Core Affect Model belongs to the group of psychological constructionist models that suggest that an emotion is constructed out of more basic components (Russell, 2003; Barrett, 2006). Core affect can be defined as “a neurological state that is consciously accessible as a simple, nonreflective feeling that is an integral blend of hedonic (pleasure-displeasure) and arousal (sleepy-activated) values” (Russell, 2003: 147). In other words, core affect “is the ongoing, ever-changing state that is available to be categorized during emotion conceptualization” (Barrett, 2006: 31), which reflects the similarity between the development of emotion concepts and other concepts discussed in the framework of cognitive semantics, and also suggests that emotion construction is essentially related to the on-line process of meaning construction. Such categorization of core affect will prove to be one of the cornerstones of the Conceptual Act Theory of Emotion discussed in subsection 2.3.3.

The notion of “core” in core affect has multiple implications (Barrett, 2006). First of all, “core” can refer to some basic sort of “core” knowledge, and it is meant to distinguish core affect form the general notion of affect. Secondly, “this form of affective responding is “core” because it is influenced by a very simple form of meaning analysis” (Barrett, 2006: 31), regardless of the value assigned to the stimulus. Additionally, “core” also means that the “basic instrumental behaviors [...] are part of this form of affective responding” (ibid.: 31). Finally, it is suggested that “core affect is universal to all humans [...] and forms the core of emotion experience” (Duncan & Barrett, 2007: 1186). Furthermore, the influences of various stimuli “allow core affect to always be represented verbally as feelings of valence, and it is often represented as arousal” (Barrett, 2006: 31). This in turn suggests that the dominant perception of core affect appears in the form of valence (e.g. positive or negative), while the degree of activation, i.e. arousal is often more difficult to capture through the use of self-report measurements.

Russell (2003) also suggests that core affect can be attributed to a stimulus, but can also appear in a free-floating form, unrelated to any particular stimulus (e.g. in moods). Another important fact pertains to the difference between core affect that is related to an individual experiencing it, and the affective quality which is situated in the stimulus. Namely, “objects, events and places (real, imagined, or anticipated) enter consciousness affectively interpreted” (Russell, 2003: 149), and the perception of affective quality influences


individuals’ reactions to them. This in turn leads to the notion of emotional meta-experience that Russell (2003) uses to describe an individual’s experience of a particular emotion. This represents “a categorization of one’s state” (ibid.: 150), or in other words, “categorizing affect puts a person into a state that corresponds with the colloquial idea of “having an emotion”” (Barrett, 2006: 21). Additionally, it is important to note that Russell (2003) places special emphasis on prototypical conceptual representations.

2.3.2.3.1. Core Affect and Cognition Building on the results from the corpus of contemporary research, Duncan and Barrett (2007) stress

the idea that affect and cognition are not two completely separable mental processes, as had been stipulated in the traditional psychological frameworks, but are instead largely connected and tend to interact with each other. Such line of reasoning lends support to the claim that “affect is a form of cognition” (Duncan & Barrett, 2007: 1185), which is further reinforced by the finding that “the circuitry that instantiates a core affective state is widely distributed throughout the brain, and includes so-called “cognitive” areas” (ibid.: 1201), i.e. “no brain areas can be designated specifically as “cognitive” or “affective”” (ibid.: 1187). Additionally, the authors propose that, depending on the focus of attention, core affect can appear as either the central or a background trait of consciousness. In the latter case, “backgrounded core affect is experienced as a property of the external world, rather than as the person’s reaction to it” (ibid.: 1202), whereas in the former case, when it is foregrounded, “core affect will be experienced directly as pleasant or unpleasant content with some degree of arousal [...] and can serve as information for making explicit judgments and decisions” (ibid.: 1203).

The multifaceted value of core affect is reflected in the fact that it “makes external information from the world personally relevant to people, providing them with a first-person experience of the world, a fluency with language so that they can describe those experiences, and enhances how those experiences are encoded for future use” (ibid.: 1196-1197). In these terms, core affect can be argued to give ‘color’ and more specific shapes to the human conscious awareness of the world. Bearing in mind that “all words have an affective dimension of meaning” (ibid.: 1198), it is clear why core affect has an invaluable role in language. Furthermore, “the affective dimension in language makes communication personal and easy to accomplish, and is an important contributor to language fluency” (Altariba & Mathis, 1997, cited in Duncan & Barrett, 2007: 1199). In that sense, the abilities to both extract the affective meaning of words, and successfully communicate that affective meaning, are the underlying requirements for language fluency.

Apart from linguistic messages, affective meaning is also present in the non-linguistic forms of communication. In addition to its role in language, core affect can also be suggested to influence the formation of background knowledge structures, in the sense that it “helps determine which experiences of the world are encoded in the brain for later use” (Duncan & Barrett, 2007: 1200). This line of reasoning also reflects the ideas of embodied cognition, where the human cognitive architecture is understood to be conditioned by the


various types of sensory experiences (e.g. Saeed, 2003; Lakoff & Johnson, 2003[1980b]; Gibbs & Colston, 2006[1995]) which in turn have an inescapable affective dimension.

In summary, affect cannot be divorced from cognition, since the specific neural circuitries that support the two phenomena largely overlap (Duncan & Barrett, 2007). Consequently, the process of meaning construction coincides with the process of emotion construction and core affect categorization, so that one constantly influences the other. While core affect will necessarily influence the on-line computations of meaning, these on-line computations, along with the context that is introduced, will also influence the construction of a specific emotional experience that takes place within that same context. Such development reflects the dynamics of the human mind, where there are no explicitly separated modules reserved for individual processes. Instead, there is a constant interplay between different neural circuitries, parts of which are often shared by different processes, all of which contributes to the active, on-line construction of human experience in its fully developed form. In other words, the specific neural mechanisms, coupled with the real-world contextual elements and background knowledge, can be understood as the grounding that sanctions the completely elaborated experience at the human scale.

2.3.3. The Conceptual Act Theory of Emotion

Like the previously discussed Core Affect Model, the Conceptual Act Theory of Emotion also “belongs to a class of models known as psychological constructionist models, which view emotions as psychological events that are composed of more basic psychological components” (Lindquist & Barrett, 2008: 902). In other words, “emotions are not biologically given, but are constructed via the process of categorization” (Barrett, 2006: 27), during which more abstract affective experiences, dubbed core affect, are conceptualized as more specific emotional experiences. As a result, “the experience of emotion does not issue from discrete bursts of activity in putative emotion mechanisms, but rather from an act of conceptualization” (ibid.: 38). This in turn reflects the previously introduced idea that emotions are not entities with strictly defined boundaries, but rather fuzzy sets without clear-cut distinctions (Russell, 1980; 2003; Niedenthal, 2008), i.e. “emotion concepts do not have conceptual cores but are represented by loose collections of situated conceptualizations”34 (Wilson-Mendenhall et al., 2011: 1107).

Namely, “concepts are not typically processed in isolation but are typically situated in background settings, events, and introspections” (Barsalou, 2009: 1238), and in that sense, situated conceptualizations refer to situated representations of categories. Bearing in mind the discussion presented in Barsalou (2009: 1283-1284), situated conceptualizations can be understood as highly specified partial frames tailored to fit specific situations. Such conclusion stems from the fact that “because a single general conceptualization would be too vague to support relevant inferences in specific situations, representations that are more

34 For a detailed account of the notion of “situated conceptualizations” see Barsalou, 2009.


specialized are constructed instead” (ibid.: 1283). Therefore, “a concept is the skill or ability to produce a wide variety of situated conceptualizations that support goal achievement in specific contexts” (ibid.: 1283). Consequently, the Conceptual Act Theory proposes that “emotion concepts [...] refer to entire situations, and thereby represent settings, agents, objects, actions, events, introspections, and mentalizing” (Wilson-Mendenhall et al., 2011: 1107), which again suggests that emotion concepts include the recruitment of partial frame structures, i.e. situated conceptualizations, and the entire process of emotion construction is similar to the active construction of a mental space that takes place during the on-line semantic processing. With that in mind, the two main ingredients of the Conceptual Act Theory are core affect, and the conceptual knowledge about emotions.

Another important parameter foregrounded by the Conceptual Act Theory is the importance of individuals’ previous experience, i.e. their background knowledge related to specific emotion concepts, which serves to both constrain and direct the categorization of core affect. In addition to background knowledge, the effects of the immediate context are not to be disregarded either, since each conceptualization of a particular emotion “is situated, in that it is a highly specialized package of conceptual knowledge that is tailored to the needs of the person in a given context” (Barrett, 2006: 33), i.e. “knowledge about emotion is context-dependent [...] and driven by emotion language” (ibid.: 38). Coupled with the fact that “language plays a strong causal role in conceptual development of emotion knowledge” (ibid.: 34), the previous stipulations suggest that the knowledge about emotions is acquired along the same guidelines as the knowledge about other concepts, with language as a conduit. In other words, “emotion concepts are abstract concepts that work in fundamentally the same way as other kinds of abstract concepts” (Wilson-Mendenhall et al., 2011: 1107), and the Conceptual Act Theory “suggests an intrinsic role for language in the emergence of emotional experience” (Barrett, 2006: 37).

An important outcome of the present model resides in the fact that the modulations of core affect during the process of emotion construction can influence cognition and behavior (Barrett, 2006; Duncan & Barrett, 2007), and in that sense, “emotion categories can be thought as goal-directed categories that develop to guide action” (Barrett, 2006: 37). Additionally, “conceptual knowledge about emotion constitutes expertise about how to deal with your own internal state – experienced as “an emotion” – and the situation or event that you believe caused that emotion in the first place” (ibid.: 36). Consequently, this suggests that the categorization of core affect also influences an individual’s reasoning regarding future actions within the given context. In other words, the frame associated to a certain emotion concept carries in it a background-knowledge-based formula that ‘tells’ an individual how to express the on-line computed emotion depending on the context. For example, if an individual experiences fear before a job interview, he should try to compose himself by, for example, controlling his breathing. If, on the other hand, an individual experiences fear during an encounter with a wolf while taking a walk in the forest, the right course of action would be to flee the scene, or climb the nearest tree. In summary, the presented course of reasoning supports the initial claim that


emotional involvement can influence cognition and behavior, and it also reinforces the role of background knowledge and immediate context in this process.

In addition to its purely theoretical appeal, the Conceptual Act Theory of Emotion has also been supported by empirical research. The first study that provided empirical support for the present theoretical framework was conducted by Lindquist and Barrett (2008), where the two authors investigated the world-focused experience of fear as a conceptual act. Their findings showed that “the world-focused experience of fear can be produced by the interplay of two more basic psychological ingredients: core affect and conceptual knowledge of emotion” (Lindquist & Barrett, 2008: 902). Wilson-Mendenhall, Barrett, Simmons, and Barsalou (2011) conducted another empirical study, where the following two main hypotheses, derived from the Conceptual Act Theory of Emotion, were tested: 1) “emotion is constructed differently depending on the situation” (Wilson-Mendenhall et al., 2011: 1120), and 2) “the composition of a situated conceptualization for an emotion would draw on contributions from different sources of information in the distributed neural circuitry that produces emotion in general” (ibid.: 1111). The results of the experiment supported both of the two hypotheses.35

Finally, it is important to stress the major similarities and differences between the Core Affect Model, and the Conceptual Act Theory of Emotion, bearing in mind that the two frameworks are largely connected. Namely, as discussed in Barrett (2006: 41), both models deal with continuous changes of core affect; both models claim that the changes in the body need not correspond to specific templates for discrete emotions; and the fact that core affect can undergo additional meaning analysis is also supported by both frameworks. On the other hand, unlike the Core Affect Model, the Conceptual Act Theory proposes that categorization plays a much broader role in emotion construction, in the sense that “even nonprototypical conceptual representations will be used to conceptualize core affect as emotional when the situation demands it” (ibid.); the Conceptual Act Theory places stronger emphasis on more contextual and socially situated emotion concepts; and finally, in the Conceptual Act Theory framework, linking an object to core affect is viewed as a consequence of categorization, whereas the Core Affect Model sees this as a purely attributional process. 2.3.4. Conceptual Act Theory of Emotion and Conceptual Blending Theory: Points of Intersection

With the previous subsections in mind, it is transparent that the presently discussed framework can be directly linked to some of the ground-stone-ideas of cognitive semantics and cognitive linguistics in general. First and foremost, the similarities between the two approaches include the emphasis on the processes of categorization and conceptualization which provide the scaffolding for both meaning construction and emotion construction. In that sense, the categorization of more abstract core affect along with the recruitment of

35 For a detailed account of the two experimental studies consult the respective references. Being beyond the scope of the present research, this section lists only the most important conclusions from the two studies.


background knowledge associated to more specific emotion concepts can also be understood as analogous to the process of meaning construction. Another valuable point of intersection is the importance attributed to background knowledge and the immediate context, which impose a significant influence on both meaning construction and emotion conceptualization. Furthermore, the previously stipulated link between language and emotion knowledge renders the two phenomena directly connected, which in turn licenses the claim that the process of meaning construction is in fact interwoven with the process of emotion construction. This idea is further supported by the empirical findings from contemporary research presented in Duncan and Barrett (2007).

When placed in the context of a linguistic stimulus-response environment, which will be discussed in the second part of the present research, it can be hypothesized that the link between the processes of meaning construction and emotion construction can be further elaborated as a cause-effect relation, where the two processes take place parallel to each other, and can directly influence one another. In short, the computed meaning, conditioned by background knowledge and the immediate context, and the specific emotional reaction, interact to sculpt each other’s final shapes. Such sculpting will necessarily involve the process of core affect categorization that shares the same context as the process of meaning construction, and may involve the recruitment of additional background knowledge structures related to specific emotion concepts. The interplay of background knowledge structures related to the frames involved in meaning construction, and background knowledge from the frames related to particular emotion concepts, coupled with the influence of the immediate context, will cause the experience of a specific emotion to vary, a proposal which is in line with the general arguments presented in Barrett (2006), Wilson-Mendenhall et al. (2011), Russell (2003), and Barsalou (2009).

In addition to the already stipulated link between situated conceptualizations and highly specified partial frame structures, a connection can also be established between background knowledge structures available for recruitment during the on-line cognitive work and entrenched situated conceptualizations. Namely, “over time, the situated conceptualization becomes so well established that it becomes active automatically and immediately when the situation arises” (Barsalou, 2009: 1284), just like familiar background knowledge structures get recruited in the process of meaning construction. Furthermore, Barsalou (2009) proposes that “once situated conceptualizations become entrenched in memory, they support a pattern completion inference process” (1284). This is reminiscent of the process of completion involved in the development of emergent structure in the context of the CBT, where the structure of the organizing frame of the blend is completed by additional entrenched background knowledge structures related to the frame in question. Viewed in the context of situated conceptualizations, “when a partially viewed situation activates a situated conceptualization, the conceptualization completes the pattern that the situation suggests” (ibid.: 1284).


2.3.5. Affect, Emotions, Media, and Politics

The mass media are a powerful tool for inducing and altering individuals’ emotional states (Barlett & Gentile, 2011; Lang & Ewoldsen, 2011), which is often exploited for purposes of persuasion, especially in the context of political discourse (Schemer, 2012; Ridout & Searles, 2011). As a result, “political scientists have recognized that both substantive message content and emotional content are important in understanding the effects of campaign advertising on voters” (Ridout & Searles, 2011: 440). Namely, placing a person in a desired state of mind can influence their decision-making and behavior, owing to the fact that “affect will influence one’s information-processing abilities” (Barlett & Gentile, 2011: 62). Additionally, both the perception and the emotional experience in individuals “have been shown to have profound effects on many aspects of on-line message processing and subsequent related behavior” (Lang & Ewoldsen, 2011: 79).

In his overview of the previous research dealing with emotional responses to political campaigns, Schemer (2012: 415) concludes that “political campaign communication clearly aims at voters’ emotions.” The author also argues that one of the main benefits of voters’ emotional and affective involvement is situated in the fact that such involvement can be used to direct their attention. For example, as shown in previous research, “attack ads aroused voters’ negative emotions thereby influencing their candidate evaluations” (ibid.: 413). Consequently, “on one hand, affective reactions are precursors to the attention to political advertising. On the other hand, the attention to such ads elicits affective reactions” (ibid.: 414), suggesting that one process fuels the other. The author also proposes that negative affective reactions are especially suitable as attention-directing devices. Additionally, in his study Schemer (2012) “demonstrated the reciprocal relation between attention to political ads and negative affective reactions in political campaigns” (ibid.: 426), where one amplifies the other. Namely, the author showed that negative political ads against asylum seekers in Switzerland produced negative affective reactions against that entire group, and these negative reactions were further reinforced due to the attention that they provoked in the first place.

Ridout and Searles (2011) conducted a study in which they analyzed the conditions under which emotional appeals are exploited in political campaigns, and which types of appeals are favored by politicians. They opened their discussion with the claim that “there is little question that political campaigns make appeals to voters’ emotions” (ibid.: 439). Based on the examination of “631 unique ads aired in 26 different U.S. Senate races from the year 2004” (ibid.: 446), the authors managed to conclude that there are certain systematic patterns that include the use of specific emotional appeals during the course of a political campaign. For example, the use of anger “is much more predictable than the use of other appeals, such as pride or enthusiasm” (ibid.: 455). Additionally, despite the obvious manipulative dimensions of the use of emotional appeal in the political setting, the authors also stress the importance of context, and the stage of the political race, which serve as two very important variables that determine the use of one type of the emotional appeal in favor of the other. Finally, although Ridout and Searles’ study stresses the importance of an emotional


appeal in TV campaign ads, the importance of provoking an emotional appeal in press reports should not be neglected either. 2.3.6. Section Summary

In brief, the present section gave an overview of some of the dominant models of affect and emotions that provide the grounds for practical measurements of real-time responses to specific stimuli. Namely, this section introduced the two dominant dimensional models of affect – the Circumplex Model of Affect and the PANAS model, which was followed by the discussion of the Core Affect model and the Conceptual Act Theory of Emotion. Additionally, some results of contemporary research in political psychology, as well as the relationship between affect, emotions, media, and politics have also been presented.

Previous Research 59

3. Previous Research

In addition to the previous research already discussed within the theoretical framework of the paper, the current section will present a few additional research papers dealing with metaphors and conceptual blending in political discourse.

3.1. CMT Framework

Semino and Masci (1996) discussed conceptual metaphors in the discourse of Silvio Berlusconi, and found that the most prominent metaphors used by this politician were those “drawing on the source domains of football, war and biblical images” (Semino & Masci, 1996: 245). The quantitative analysis showed that the football metaphors were the most prominent and most frequent, followed by the war metaphors. The fact that Berlusconi created “the whole identity of his new party around the football metaphor” (ibid.: 250) enabled him to both bring his supporters closer together by aiming at their emotional attachment to the newly-founded political team, and to bring them closer to the complex political topics through the use of simple sport terminology. Additionally, the analysis revealed that “Berlusconi ... frequently mixed the football and war metaphors” (ibid.: 253) which led to a “metaphorical system that is likely to sound natural and coherent” (ibid.: 254), owing to the fact that both metaphors are pervasive in political discourse, and can be viewed as an extension of conventional metaphors SPORT IS WAR and WAR IS SPORT (ibid., 1996).

Burnes (2011), analyzed metaphors used in French and British newspaper reports of the 2008 elections that took place in Pakistan and the US. The study showed that the most frequently used metaphors were those of conflict. Namely, “in the US election reports, conflict metaphors constituted almost a quarter of metaphors (24%) describing the elections, and in the Pakistan reports, they comprised over a third” (Burnes, 2011: 2166). Among others, the corpus analysis also revealed the presence of ‘sport’ and ‘journey’ metaphors. With “the reciprocal relationship between WAR and SPORT metaphors” (ibid.: 2169) in mind, and taking into consideration that “as source domains, sports and war both equate physical power with political power” (ibid.: 2169), it is understandable why the two domains are so suitable for use in the political arena. Additionally, such results suggest that “Western society places a positive value on conflict, and conceptualizes politics primarily in terms of conflict” (ibid.: 2174).

Journey metaphors were especially prominent in reports of US elections, where Obama was portrayed as an emissary “sent by Martin Luther King to continue the journey to a better America” (ibid.: 2171). Furthermore, “Obama used the journey metaphor in his victory speech in a way which echoed Martin Luther King” (ibid.: 2170). Such framing of political ideas is aimed at obtaining positive evaluations from the public through the emphasis on equality and the pursuit of a better future. This gives another illustration of the potential for highlighting and hiding that the conceptual metaphor offers.


3.2. CBT Framework

Coulson and Oakley (2006) analyzed two examples of persuasive discourse to show how both cases “use discourse to prompt very specific actions in the world” (ibid.: 48), and how the CBT framework accounts for the mental operations that license such real-time influence of discourse onto people’s behavior. Still, the authors emphasize that “the rhetorical effect of the text [...] depends in part on the reader’s willingness to construct the blend” (ibid.: 51). The two examples include “a widely distributed email message urging recipients to vote for Democratic candidates in the 1998 U.S. congressional election” (ibid.: 47), sent by Michael Moore, and “a solicitation for monetary donations from the St. Matthew’s Church Ministry” (ibid.: 47). The present section will address only the most interesting details related to the first example, where five distinct instances of conceptual blending that can be found in Moore’s argument are presented. These include palatable candidates, stinky candidates, public conversation, sending a message, and the legal act of civil

disobedience. If you want Congress to stop this witch hunt, if you want Congress to start focusing on the REAL problems facing this country and the world ... get out and vote November 3. Hold your nose if you have to. (ibid.: 51)

For example, the above excerpt illustrates the stinky candidates blend which involves input spaces of voting and holding one’s nose while performing an action. The voting frame recruits a familiar script of going to the ballot and casting a vote for one of the candidates, whereas the organizing frame of the other input involves an unpleasant action. As the process of conceptual integration progresses, voting is framed “as an unpleasant but necessary chore” (ibid.: 51), and the entrenched nature “of the ‘stinks’ metaphor allows speakers to understand the text as acknowledging the limited political options available to progressive voters” (ibid.: 51). Consequently, in the blend “the voter is performing an unpleasant task in a stench-ridden environment, and that task is to choose the thing that stinks the least” (ibid.: 52). Such integration network is expected to prompt the reader to vote for the candidate that stinks less, i.e. for the Democrat.

In the discussion of the public conversation blend, the authors suggest that it “has desirable rhetorical characteristics from both a cognitive and an affective standpoint” (ibid.: 53). This suggests that the blend simplifies the course of events, and “if the reader fully integrates knowledge about the political process with her own personal experience with losing arguments, it can evoke the sorts of emotions that accompany the latter” (ibid.: 53). In turn, such line of reasoning supports the general idea that conceptual blending can serve to modulate affective states, which will be pursued in sections 4.1 and 4.2 of the present paper.

Coulson and Oakley (2008) analyzed the CONNECT THE DOTS metaphor from a radio interview with the former White House Special Advisor on Counterterrorism, Richard A. Clarke. One of the authors’ goals was to attempt to integrate the CBT framework and the Mental Space Theory with “Chafe’s (1994) theory of discourse based on conscious experience” (Coulson & Oakley, 2008: 28). Additionally, the authors argued that the understanding of the excerpts from the above mentioned interview “relies upon three interrelated kinds of


knowledge that we call elements of understanding” (ibid.: 28). These elements include linguistic knowledge, cultural knowledge and situational knowledge.

The situational knowledge that Coulson and Oakley discuss includes the necessary information about the setting, participants, and the immediate circumstances that underlie the given event. In this particular example, the participants are Richard Clarke, Dave Davies, the journalist, and Richard Posner, a virtual participant (ibid.). The setting is constrained by the radio interview, while the “rhetorical context of the exchange is the differing assessment of Clarke (the actual guest) and Posner (the virtual participant) of the Congressional report by the 9/11 Commission” (ibid.: 32).

The cultural knowledge assumes the awareness that the CONNECT THE DOTS metaphor is largely related to the events from 9/11, i.e. the terrorist attack on the World Trade Center. In that sense, “failure to connect the dots refers to the failure by US intelligence agents to understand the relationship between different facts about individuals with links to terrorist groups” (ibid.: 35). In other words, the important information that was unknown to the agencies corresponds to the unrecognized pattern in the game.

In addition to the familiarity with the socio-cultural context, Coulson and Oakley (2008) also suggest that the success of this metaphor is largely licensed by the linguistic knowledge of the meaning of the concept connect. With that in mind, they suggest a link between the CONNECT THE DOTS metaphor and the entrenched metaphor knowing is seeing. Consequently, Colson and Oakley (2008: 36) propose that “the connect the dots blend co-opts the KNOWING IS SEEING metaphor and applies it to the game of connect the dots by adding a few new mappings.” Additionally, such line of reasoning supports the authors’ idea that “creative metaphoric blends rely heavily on conventional linguistic knowledge” (ibid.: 46).

The conceptual integration network that the authors propose in the first part of their analysis (ibid.: 31-37) consists of a presentation space with the “connect the dots” game, the reference space, with national security agencies and the information related to them, and the blend, where “it is possible for the intelligence officer to see terrorist activities represented in the dots” (ibid.: 36-37). An important factor to be taken into account is that not all dots in the blend have the same relevance for solving the puzzle, and that the agencies’ failing to connect the dots might not be caused by the failure to share information, but by the lack of time to go over the potentially relevant links within the available information (ibid.: 41). Coulson and Oakley also stress the fact that the interpretation of the metaphor CONNECT THE DOTS will be conditioned by the individual’s linguistic and cultural knowledge, as well as by the immediate context and the level of involvement that the individual ‘invests’ into (de)constructing the blend.

Additionally, the authors also discuss the SCREAMING DOTS metaphor which builds on the initial conventional CONNECT THE DOTS mapping, and also introduces new elements of fictive interaction. Furthermore, the authors claim that in the case of the SCREAMING DOTS metaphor “innovation [...] emerges from the intersection of cultural, linguistic, and situational knowledge in the service of the interconnected demands of discourse coherence and argumentative goals” (ibid.: 47). Namely, because Clark disagrees with the


conclusions arising from the CONNECT THE DOTS integration network, “he explicitly disputes the applicability of the term connect” (ibid.: 47), and employs another conventional blend available for recruitment – that of fictive interaction. For a full account of this part of the analysis, consult Coulson and Oakley, 2008: 41-47.

Harder (2005) discusses the notion of full-blown blending, where this process “ is distinguished by the way it enables the conceptualizer to superimpose two pictures that could not occur together in the reality space – without ‘cheating’, i.e. in a way that preserves the contradictions while suspending them for the purposes of creating the blended space” (ibid.: 1650). In layman’s terms, full-blown blending enables an individual to construct a blended space that contains elements in direct opposition to some elements from the reality spaces, which is done for purposes of persuasion, or even self-preservation. Still, the individual does not do away with those contradictions, but keeps them active. In that sense, Harder (2005) argues that, when blends are viewed in their functional context, “a distinction becomes necessary between the purely conceptual anatomy of a blend and the status of a blend as a successful response to a particularly difficult challenge” (ibid.: 1640). In relation to that line of reasoning, another important variable that Harder (2005) places into the equation is the role of the social grounding, i.e. the immediate socio-political setting, which further stresses the functional perspective of the cognitive processes by showing how they are “shaped by and contribute to the pattern of interaction of which they constitute one aspect” (ibid.: 1637). Additionally, Harder (2005) also proposes that apart from the processes of integration, attention should also be paid to the processes of disintegration, since, under certain circumstances, conceptual integration networks can become unstable and disintegrate.

This author also introduces the notion of polarization which he illustrates in relation to post-9/11 conceptualizations of American policies directed at the Muslim world reported in the news and commentaries. Namely, polarization takes place “when two parties interact in such a way that each act in a sequence brings about a greater sense of threat in the other party” (ibid.: 1643). In other words, framing the opposition as a threat serves to both fuel and justify the negative campaign against them. This was particularly evident in the post-9/11 period, when the U.S. framed almost the entire Muslim world as the enemy, and presented its foreign policy as war on terror.

Interviewer: Is there not a risk of alienating the Muslim world and thus sowing the seeds of future terrorist activities? U.S. official: You are playing on Bin Laden’s turf. I think you should stop playing on his turf. (ibid.: 1646)

Although from a purely objective point of view the constructed mental space in the above example may seem inadequate, from the policymaking point of view it may be purely legitimate. Namely, the U.S. official imposes a direct constraint on the entire line of reasoning by enforcing a favorable point of view, i.e. a point of view that justifies the U.S. military intervention in Iraq, and at the same time secures the safety of the U.S. Such course of action reflects the cognitive, emotional, and functional mechanisms of conceptual (dis)integration that Harder (2005) discusses in his paper. Additionally, Harder (2005) pays particular attention to the emotional dimension of the process of conceptual integration, in the sense that “getting an ‘integrated’ picture of objects, where fragmentary experiences with different emotional colorings are brought together, is an emotional


achievement as well as a cognitive achievement” (ibid.: 1642). The remainder of Harder’s (2005) paper engages in an in-depth analysis of full-blown blending and disintegration that took place when George W. Bush, president of the U.S. called Ariel Sharon a “man of piece.”36

3.3. Political Psychology

Bougher (2012) stresses the fact that the majority of previous research dealing with the use of conceptual metaphor in political discourse was focused primarily on the elite discourse, while the metaphors used by the citizens who comprise the electorate have been largely neglected. The author foregrounds the importance of these metaphors, since an in-depth investigation in this field could provide a more through insight into the minds of the voters and the ways they conceptualize the political scene, which “holds relevance for a number of other areas within the field of political psychology” (Bougher, 2012: 154). Additionally, such investigation could show which metaphors the electorate responds to best, as well as whether the citizens employ similar metaphors to those of the elite. Accordingly, Bougher’s paper also emphasizes that “a focus on elite discourse risks understanding the cognitive depth citizens apply in political reasoning and the consistency that binds their beliefs and attitudes” (Bougher, 2012: 151). In relation to the use of specific metaphors, the author discusses the role of analogical reasoning, as well as issues of similarity and familiarity, which condition the choice of domains, i.e. inputs, involved in metaphorical mappings.

Furthermore, the author also discusses the emotional appeal of conceptual metaphors, and suggests that “metaphorical reasoning can help illuminate the “rationale” of affect in political reasoning” (ibid.: 155). Additionally, based on the overview of previous research, Bougher (2012) concludes that “identifying the metaphoric sources used in political reasoning cannot only clarify the origins of those emotions; it can also demonstrate the cognitive complexity that underlies emotion and intuitive understanding” (ibid.: 155). Finally, bearing in mind that “metaphor offers a cognitive mechanism that explains how citizens make sense of the political world by drawing from their nonpolitical knowledge and experiences” (ibid.: 157), a thorough investigation of metaphors used by both the political elites and the electorate can provide a comprehensive insight into the ways in which the political reality is constructed, understood, and navigated.

3.4. Section Summary

Bearing in mind both the previous research discussed in the present section, and the previous research presented in the theoretical framework, it can be concluded that metaphor appears as a ubiquitous phenomenon in political discourse. Such frequency of metaphor use in political discourse is justified mainly by its role in persuasion, predominantly achieved by constructing a favorable point of view, and by provoking an 36 For a detailed overview of this part of the research consult Harder, 2005: 1646-1650.


emotional appeal with the electorate. The current section provided a brief overview of some results from contemporary research conducted in the domain of the CMT, CBT, and political psychology.

Present Research 65

4. Present Research

The present research will be conducted in two stages. Namely, the first stage of the present research will involve the theoretical analysis of the corpus, following the methodology presented in section 4.1. The second stage of this research will be focused on investigating whether metaphorical expressions commonly used in the political discourse of daily newspapers can in fact provoke a tangible emotional reaction with the readers. Specific methodological procedures employed in this part of the research are elaborated in section 4.2. With such an outline of the course of analysis in mind, the present research will attempt to answer the following Research Questions:

[1] What are the basic functions of conceptual metaphors in the political discourse of daily newspapers?

[2] How are these functions achieved, i.e. which mechanisms are they licensed by (for example, are there mechanisms like emotional appeal, highlighting and hiding, simplification, etc.)?

[3] Which groups of metaphors (based on their conceptual keys) are dominant, and are the findings consistent with the results from previous research?

[4] What type of networks is dominant: single-scope, or double-scope?

[5] Which vital relations are dominant in metaphorical networks, and what happens to them during compression?

[6] Is there any similarity between the forms of emergent structures that appear in the present corpus?

[7] Can conceptual metaphors from the political discourse of daily newspapers provoke an actual, tangible emotional response with the audience, i.e. readers?

[8] How was that emotional response depicted in terms of affect and more specific emotion concepts? In particular, the first six research questions will be addressed in the first part of the research, while the final two research questions will be the subject of inquiry of the second part of the present research. Additionally, research questions 1 through 3 fall in the domain of the CMT, research questions 4 through 6 will be dealt with in the framework of the CBT, while research questions 7 and 8 will be addressed within the theoretical framework dealing with affect and emotions.

Present Research: Part 1 66

4.1. Present Research: Part 1

The first part of the present research will be dealing with the theoretical analysis of the metaphorical expressions from the corpus. Corpus formation criteria and steps involved in the analysis, as well as the main theoretical frameworks that are employed, are described below.

4.1.1. Corpus and Methodology

The corpus of the paper consists of metaphorical expressions from columns dealing with foreign and domestic politics, extracted from the online editions of the daily newspaper The New York Times in the period from 28 November 2011 to 31 December 2011, and there are a total of 240 metaphorical expressions. The criterion for Metaphor Identification was adopted from Charteris-Black (2004), where conceptual metaphors are defined in terms of semantic tension, role in persuasion, and cross-domain mappings in the conceptual system. In that sense, the CMT framework played a vital role in corpus construction. The Metaphor

Identification process consisted of two parts: the first part was a close reading of the articles to identify conceptual metaphors and metaphor keywords which represent words often used with a metaphoric sense, “and it is possible to measure the presence of such keywords quantitatively in the corpus” (Charteris-Black, 2004: 35); in the second part, the article context was used to evaluate whether the use of a keyword was literal or metaphorical (Charteris-Black, 2004). Metaphor Identification was then followed by Metaphor

Interpretation and Metaphor Explanation, where a link was established between the metaphorical expressions and cognitive factors they are governed by, and the role that metaphors perform in political discourse was further explored (Charteris-Black, 2004).

The analysis then follows the basic guidelines as presented in Cameron and Low (1999). Namely the collected metaphorical expressions were grouped according to the general metaphors they represent, and finally the results were used to “suggest understanding or thought patterns which construct or constrain people’s beliefs and actions” (Cameron & Low, 1999, cited in Charteris-Black, 2004: 34). In line with such an approach, metaphorical expressions were classified according to their corresponding conceptual metaphors, which are then grouped under the conceptual keys (in the sense of Charteris-Black, 2004) POLITICS IS CONFLICT

and POLITICS IS SPORT. Such corpus choice is justified by the fact that newspapers can be considered a secondary carrier of

political discourse. Namely, daily newspapers are aimed at a wide audience of various backgrounds, education, and professional orientation. Therefore, newspaper language can be considered a good approximation of everyday language, which serves the purpose of presenting to the public the ideas and goals of certain political groups. In that sense, newspapers serve as a mediator between political fractions and citizens and perform the role of the secondary carrier of the political discourse.


Once the corpus was formed, the analysis entered the domain of the CBT. Namely, individual examples from the corpus were analyzed through the steps presented in Table 1. Such an approach was expected to provide a detailed insight into the structure of metaphorical conceptual integration networks, by addressing some of the key tenets of the CBT framework. Additionally, the algorithm for the analysis given in Table 1 should yield systematic results that are expected to produce statistically relevant data based on the present corpus. This part of the analysis also tackled the problem of functions that metaphors play in political discourse, as well as some of the mechanisms through which these functions are achieved, with the potential to provoke an emotional appeal as one of the most salient of these mechanisms. With this in mind, the CMT framework was also included in the analyses.

Step 1: Identify the inputs, the generic space, and their elements.

Step 2: Identify counterpart connections and partial mappings.

Step 3: Identify the organizing frame(s), and the type of the integration network (simplex, mirror, single-scope, or double-scope).

Step 4: Identify the emergent structure in the blend, and the processes of composition, completion, and elaboration (coupled or decoupled).

Step 5: Identify possible backward projections and their consequences.

Step 6: Identify Vital Relations and their compressions (i.e. compressions of outer-space vital relations into inner-space vital relations).

Step 7: Identify the Governing Principles for Compression that have been satisfied.

Step 8: Discuss the systematic nature of compression (in the sense of Fauconnier & Turner, 2008).

Step 9: Discuss the issue of network optimization.37

Table 1. Main steps in corpus analysis within the CBT framework

4.1.2. Corpus Analysis and Discussion

The present section will explore the structure, basic functions, and mechanisms through which these

basic functions of conceptual metaphors in the political discourse of daily newspapers are achieved. Additionally, special attention will be paid to the notions of compression, systematic nature of compression (in the sense of Fauconnier & Turner, 2008, and Fauconnier, 2008), and network optimization.

37 This step of the analysis will be addressed separately in section 4.1.3, where the issue of network optimization based on the results obtained from CONFLICT and SPORT metaphors is thoroughly investigated.

CONFLICT Metaphors 68

4.1.2.1. CONFLICT Metaphors Metaphors of CONFLICT will be organized under the conceptual key POLITICS IS CONFLICT. This will

include conceptual metaphors with their corresponding metaphorical expressions, organized around the following metaphor keywords: battle, war, fight, showdown, standoff, attack, defense, strategy, maneuvering,

victory, defeat, alliance, and (war) front. The distribution of these metaphor keywords in this part of the corpus is presented in the bar graph in Figure 12. Percentage-wise, CONFLICT metaphors take up 59.58 % of the corpus, and there is a total of 143 metaphorical expressions that can be attributed to the above mentioned conceptual key. Such a dominant role of CONFLICT metaphors is in keeping with the findings from previous research (e.g. Howe, 1988; Burnes, 2011).

Figure 12. Distribution of metaphor keywords in CONFLICT metaphors

4.1.2.1.1. Structure of CONFLICT Metaphors

The current section of the paper will use examples from the corpus to discuss the structure of metaphorical conceptual integration networks. Some examples of metaphor systems and interactions of metaphorical networks within those systems will also be addressed. (1) The battles within the movement have played out in public. (The New York Times, 3/12/2011)

This metaphorical expression corresponds to the conceptual metaphor POLITICAL ARGUMENT IS A BATTLE, which further corresponds to the conceptual key POLITICS IS CONFLICT (in the sense of Charteris-Black, 2004). The generic space contains schematic structure that corresponds to both inputs, and it is related to the event structure metaphor. Namely, it contains the following elements: participants, setting, means, and goal. The source input is that of a BATTLE, and it contains a prebuilt compression of a battle that is recruited in this network, while the target input is that of POLITICAL ARGUMENT. The elements of the two inputs are presented in Figure 13. Elements: battlefield - public arena, soldiers – politicians and party reps belonging to the same


political group, officers – leading figures from the same political group, casualties - the defeated fraction within the same political group, and weapons - arguments are connected via an outer-space vital relation of Analogy that is compressed into Uniqueness in the blend, i.e. they constitute counterpart connections between the input spaces, which reflects the process of matching.

Additionally, the blend inherits the organizing frame of the source input, which renders this network single-scope. This allows for the structure from the target input to be projected into the pre-compressed inner-space relations of the source input (Fauconnier & Turner, 2002). The blend also contains emergent structure not present in either of the input spaces: arguments can cause harm, and political defeat equals political death. The emergent structure is a result of the following processes: 1. composition: the network establishes connections between the elements in the inputs and projects them to the blend; 2. completion: the elements from the target are projected into the organizing frame of the source. Additionally, the recruitment of background knowledge related to the frame of a BATTLE also aides the process of completion; 3. elaboration: since it can affect actual political outcome and reasoning, this is an instance of coupled

elaboration.

Figure 13. Battles within the movement

This emergent structure can yield backward projections to the target input, which can in turn serve to

provoke an emotional response with the public (similar to the Regatta blend, discussed in Fauconnier & Turner, 1994; 2006[1998]; 2002). If compression is taken into consideration, it can be concluded that the emergent structure is also a direct product of compression of the outer-space vital relation of Disanalogy between the


means for achieving victory in the two input spaces. An important contextual variable that needs to be stressed is the fact that the conflict described by the present network is taking place between the members of the same political group. This can in turn serve to additionally explain the form of the emergent structure, according to which arguments can harm all of the participants.

Apart from the above mentioned compression of Analogy into Uniqueness, it is useful to address the vital relations of Cause-Effect and Intentionality. Namely, the preexistent conventional compression aligns the Cause-Effect relationships in the two inputs and projects them to the blend. For example, the causal relationship between waging a battle and winning or losing it, and the causal relationship between engaging in

a political argument and winning or losing it are compressed into Uniqueness in the blend. The Intentionality that underlies the previously discussed causal chains is also compressed into Uniqueness (i.e. into the intention of winning). In terms of the Governing Principles for Compression borrowing for compression is satisfied, since the mappings introduced by the conceptual metaphor automatically recruit pre-built compressions. Compressions of one vital relation into another have already been discussed, while scalability is also satisfied, and human scale is achieved.

In order for the analysis of compression to be complete, guidelines from Fauconnier and Turner (2008) should also be addressed. Namely, in line with their argument, and bearing in mind that the notion of borrowed compression from the source input has already been mentioned, it is obvious that the source input does not represent a fundamental concept. This means that the integration network needs to be expanded beyond the minimal four-space model in order to account for the recruited compressions, i.e. for both cobbling and

sculpting (in the sense of Fauconnier & Turner, 2008). The same can be argued for the target input, since ARGUMENT can also be represented as a compression. In that sense, such an elaborate network of conceptual integration reflects the notion of systematical compressions that enable the human mind to operate over numerous compressions and decompressions, i.e. perform blending and unpacking as necessary.

The shape of this network is captured by the diagram presented in Figure 14. As can be concluded, the systematicity of compression does not end with the battle and argument inputs. On the contrary, it can be elaborated even further. Note also that both battle and argument blends have competition as one of their input spaces. Such relatedness between the source and target input of the final network can be used to account for the prealignment of their topologies before they are projected to the final blend of the fully elaborated network.

The elaborate network from Figure 14 resembles a complex molecule from the domain of organic chemistry. This “molecule-model” is plausible on at least two levels. Firstly, it reflects the notion of the dynamic structure of conceptual integration networks, where they can change depending on the context, just as a molecule can mutate under foreign influences. Secondly, the “molecule-model” stresses the fact that the CBT is not confined to the domain of static structural analysis grounded on rigid mathematical approximations. On the contrary, the CBT deals with live, dynamic phenomena that the human thought is made out of, which undergo constant changes and mutations owing to both immediate contextual influences,


Figur

e 14.

“Mole

cule-

Mode

l”


and rhetoric and pragmatic strategies that interlocutors have at their disposal. In that sense, the “molecule-model” captures the very essence of human reasoning and cognition, and renders the role of compression in this process more transparent. Additionally, the notion of pre-built compressions available for recruitment foregrounds the role of the immediate social network that provides the grounding for human reasoning. In other words, these entrenched compressions are not instances of compressed anomalies available to a small number of individuals. Instead, they make up the vast domain of background knowledge characterized by a marked cultural influence.

(2) His [Mitt Romney’s] aides are vigorously organizing in Florida, where absentee voting begins right after the New

Year, and in states far down the line, to build a backup plan when the nominating battle becomes a delegate fight. (The New York Times, 15/12/2011)

This example illustrates the interaction between two conflict metaphors: ELECTIONS ARE A BATTLE, and

ELECTIONS ARE A FIGHT. These two conceptual metaphors are organized under the conceptual key POLITICS IS

CONFLICT. Based on Figure 15, it can be concluded that both initial conceptual integration networks actually share the same generic space which is structured by the event structure metaphor, and contains the following elements: participants, setting, means, and goal. The source input of the first metaphor is structured by the organizing frame of a BATTLE which represents a prebuilt compression available for recruitment, while the target input is structured by the frame of ELECTIONS. In the second metaphor, the organizing frame of the source input is FIGHT which also represents a prebuilt compression available for recruitment, and the organizing frame of the target input is ELECTIONS. Both blends inherit the organizing frames of their corresponding source inputs, making each of the networks single-scope. In addition to being linked by the change of focus in the discourse lattice, the two blends are connected by the vital relation of Change, forged by the immediate discourse context.

Each of the blends develops emergent structure, as the central product of compression: Blend 1: winning the nomination harms the opponent; Blend 2: good campaign harms the opponent; which is achieved through the following processes: 1. composition: during this process, connections between the inputs are established and elements are projected to the blend. Counterpart connections are established between the inputs of the two initial networks, and their elements are projected to blends 1 and 2. 2. completion: elements from the target inputs of the two initial networks are projected to the organizing frames of the blends 1 and 2, respectively. Completion also involves the recruitment of background knowledge related to the organizing frames of the two blends, which facilitates the entire process. 3. elaboration: as in all cases of conceptual metaphors in the political discourse, this is again an instance of coupled elaboration, since it can affect inferences and reasoning in real time.


Figur

e 16.

Nomi

natin

g batt

le be

come

s a de

legate

fight.


The emergent structure is the result of the Disanalogy between the means for achieving the goal between the input spaces in the two initial networks, respectively, i.e. the Disanalogy between physical

violence that can lead to casualties, and a political campaign. Such emergent structure can yield backward projections from the blends back to the inputs, which can provoke an emotional response, thereby amplifying the rhetorical power of the conceptual metaphor. Another important aspect of this network that needs to be addressed is the fact that the conflict is taking place during the primary elections in the Republican Party, i.e. all of the participants belong to the same political group. This contextual variable can also account for the form of the emergent structures in the two blends, where all participants can be harmed in the political process.

The most dominant compression is the compression of the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness, evident in the two initial networks. Namely, the causal relations between the inputs of the two initial networks are related by the vital relation of Analogy, which is compressed into Uniqueness in the blend. The inner-space vital relations of Intentionality in the inputs of the two initial networks are also analogous, and are compressed into Uniqueness in the blends 1 and 2. Again, the compression of Disanalogy leads to the creation of emergent structure, unique for the blend. In sum, the satisfied Governing Principles of Compression include compressions of one vital relation into another. Additionally, scalability and borrowing for compression are also satisfied, owing to the fact that both the human scale is achieved, and pre-built compressions are activated.

All of the inputs from the two initial networks represent compressions available for recruitment, rather than basic concepts, which reflects the systematic nature of compression. As argued in Fauconnier and Turner (2008), every conceptual integration network has parts that are entrenched, and available for recruitment (cobbling), and parts that are constructed on-line (sculpting). Like in example 1 discussed above, Figure 15 captures only the on-line active process of sculpting, whereas the entrenched part, i.e. cobbling, can be represented in an elaborated “molecule-model” network similar in structure to the one in Figure 14.

(3) Late Saturday, former Senator Bob Dole, the 1996 Republican nominee, endorsed Mitt Romney, whose

campaign is now anticipating a long and hard-fought nominating battle against Mr. Gingrich. (The New York Times, 19/12/2011)

The metaphorical expression in the previous example corresponds to the conceptual metaphor

ELECTIONS ARE A BATTLE, and to the conceptual key POLITICS IS CONFLICT. Consequently, the topologies of the source and target inputs of the corresponding conceptual integration network are structured by the organizing frames of BATTLE and ELECTIONS, respectively. The process of matching, i.e. counterpart connections between the two input spaces is presented by the horizontal lines in Figure 16. The generic space is structured by the event structure metaphor and contains the following elements: setting, participants, goal, and means. Since the blend acquires its organizing frame from the source input, this is another example of a single-scope network.


The blend develops emergent structure, which is a direct product of compression of the outer-space vital relation of Disanalogy between the means for achieving the goal in the two inputs. This structure is introduced through the following processes: 1. composition: elements of the two inputs are linked by counterpart connections, and projected to the blend; 2. completion: elements from the target input are projected into the organizing frame of the source input. Through the recruitment of additional background knowledge pertaining to the organizing frame of the source input, the blend is completed; 3. elaboration: finally, mental simulation of the blend licenses on-line real-world inferences, which suggests that this is a case of coupled elaboration. Such emergent structure can give way to backward projections to the inputs, which can further amplify the rhetorical power of a conceptual metaphor.

Figure 16. Nominating battle between Mitt Romney and Newt Gingrich

As in other metaphorical conceptual integration networks, the most dominant compression is that of

the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness. Namely, cause-effect relations structuring the topologies of the two input spaces are related by the outer-space vital relation of Analogy, which is compressed into the inner-space vital relation of Uniqueness in the blend. Intentionality underlying these cause-effect topologies is also compressed into Uniqueness in the blend. Additionally, there is also the compression of Disanalogy into Uniqueness, already mentioned above. Therefore, it is obvious that


the satisfied Governing Principles of Compression include the compression of one vital relation into another. Additionally, scalability, and borrowing for compression are also satisfied.

In terms of the systematic nature of compression, the analysis of the present example also sanctions the conclusion that neither input is a basic concept, but rather a pre-built compression available for recruitment. In that sense, the conceptual integration network depicted in Figure 16 represents the sculpting part of the more elaborate network, i.e. the part that is constructed on-line. The remainder of the fully-fledged network could be represented in the form of a “molecule-model” network, similar to the one in Figure 14.

(4) Facing a tough re-election fight in five months, Mr. Sarkozy is presenting himself as a man of experience, capable of strong leadership in a crisis. (The New York Times, 2/12/2011)

The above metaphorical expression corresponds to the conceptual metaphor RE-ELECTION CAMPAIGN IS

A FIGHT, which is further contained within the conceptual key POLITICS IS CONFLICT. The organizing frame of the source input is that of a FIGHT, whereas the organizing frame of the target input is that of a RE-ELECTION

CAMPAIGN. Counterpart connections between the elements from the source and target input are presented by the horizontal lines in Figure 17. The generic space contains the following elements: setting, participants, goal,

and means, and it is structured by the event structure metaphor. The blend inherits the organizing frame of the source input, suggesting that this is a single-scope network. In this case, the conflict is taking place between representatives of different political parties.

Figure 17. Sarkozy’s re-election fight

Compression in the network sanctions the creation of emergent structure. More specifically, it is the

compression of the outer-space vital relation of Disanalogy between the means for achieving the goal in the


two inputs that leads to the emergent structure in the blend. Additionally, emergent structure is constructed through the following processes: 1. composition: counterpart connections between elements from the two inputs are established and elements are projected to the blend; 2. completion: elements from the target input are projected into the organizing frame of the blend, which was adopted from the source input space. This way, topology of the blend is completed, which is also facilitated by the recruitment of additional entrenched knowledge structures related to the FIGHT frame, i.e. the organizing frame of the blend; 3. elaboration: since mental simulation of the blend provokes inferences in real-time, this is a case of coupled

elaboration. The emergent structure can cause backward projections to the inputs which can, for example, provoke an emotional response with the readers, or make them focus only on certain aspects of the problem, which consequently augments the rhetorical force of the metaphor.

In terms of the Governing Principles of Compression, the present network satisfies the following: compression of one vital relation into another, scalability, and borrowing for compression. Namely, there are obvious compressions of the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness. Additionally, human scale is achieved, and the network builds on activated entrenched compressions, i.e. the inputs are not basic concepts, but pre-built compressions.

This idea leads to the issue of the systematic nature of compression, and to the notion of the “molecule-model.” In line with the previously presented arguments, the conceptual integration network presented in Figure 17 can be viewed as the on-line active construction, licensed by the more elaborate, fully developed network that also incorporates the pre-built, entrenched elements available for recruitment.

(5) As the perceived front-runner, he [Mr. Paul] is drawing a barrage of attacks from his rivals. (The New York

Times, 18/12/2011)

The above example represents an interaction between two conceptual metaphors: ELECTION CAMPAIGN

IS A RACE, and ELECTION CAMPAIGN IS A BATTLE, the former of which can be attributed to the conceptual key POLITICS IS SPORT, and the latter to the conceptual key POLITICS IS CONFLICT. The first integration network has a combined generic space 38 which includes elements of the event structure metaphor, and the basic preconceptual image schematic structure of the SOURCE-PATH-GOAL image schema. Namely, as a conceptual

primitive (in the sense of Mandler, 1992), the proposed image schema underlies not only the human ability to understand the world in terms of spatial relations, but also the ability to forge metaphorical connections between a concept such as movement and other concepts. Therefore, combined with the elements adopted from the event structure metaphor which also has an underlying image schematic structure, the notion of the

38 The notion of the combined generic space will be further elaborated in the section dealing with SPORT metaphors.


Figur

e 18.

Metap

hor s

ystem

: RAC

E and

ATTA

CK


combined generic space is aimed at foregrounding the importance of the preconceptual image-schematic structures in the construction of metaphorical conceptual integration networks grounded in entrenched conceptualizations, i.e. conventional conceptual keys. Still, as discussed in the theoretical framework, such notion of the generic space will be understood as a theoretical construct and an artifact of analysis.

Blend 1 adopts the organizing frame of its corresponding source input, i.e. that of a RACE, while Blend 2 adopts the organizing frame of the source input from the second network, i.e. that of an ATTACK. In that sense, both conceptual integration networks from Figure 18 are single-scope. It is also worth stressing that both the race and the conflict are taking place within the Republican Party during the primary elections. The two blended spaces also develop the central emergent structure as a direct product of compression in each of the two networks. More specifically, the two emergent structures are products of compression of the vital relations of Disanalogy that exist between the means for achieving the goal between the source and target inputs in each of the two networks, respectively. The emergent structure is constructed through the following processes: 1. composition: counterpart connections are established in each network and elements are projected to the blends 1 and 2. This process involves the compression of outer-space vital relations onto inner-space vital relations. 2. completion: each blend activates entrenched knowledge structures related to its organizing frame, which serves to complete the blend. 3. elaboration: both blends can be dubbed instances of coupled elaboration since they can affect real-time inferences and lead to actions. Such emergent structures can sanction backward projections to the corresponding inputs, which can in turn increase the rhetorical power of the two metaphors. Namely, both presenting elections as a race and as a conflict can provoke an emotional response with the readers. Furthermore, the interaction of the two metaphorical networks in the discourse lattice can render them resonators to each other, which can amplify their persuasive power even further.

Additionally, apart from being the two coinciding focus spaces in the final discourse lattice, the two blends are connected by the outer-space vital relation of Cause-Effect. Namely, the fact that Mr. Paul is the perceived front-runner invites the attacks against him. However, the two blends remain connected only via the discourse lattice, and the change of focus, and there is no final compression between them. In that sense, it could be argued that the Cause-Effect vital relation between the two blends is actually forged by the discourse structure, i.e. the immediate context. Such a conclusion could further serve to reinforce the already established invaluable role of context in the process of meaning construction argued by the cognitive semanticists (e.g. Saeed, 2003; Evans & Green, 2006; etc.).

Both networks are dominated by the compressions of the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness. Namely, both networks show the vital relation of Analogy between


the causal structures of their source – target input pairs, which is compressed into Uniqueness in the blend. The same can be said about Intentionality. In that sense, in addition to borrowing for compression, and scalability, the present two networks also satisfy the principle of compression of one vital relation into another. Additionally, as mentioned above, the outer-space vital relation of Cause-Effect between blends 1 and 2 remains uncompressed.

In view of Fauconnier and Turner’s (2008) argument on the systematicity of compression, diagrams in Figure 18 depict only the active on-line aspect of meaning construction. Capturing the activation of the massive entrenched knowledge structures and frames would require the construction of an elaborate “molecule-model” network (see Figure 14 for details). (6) As Mr. Gingrich and Mr. Romney tangled with each other over their backgrounds, their world view and their

leadership styles, the rest of the Republican field took aim at both of the men, reflecting how the race is narrowing as voters draw closer to weighing in. (The New York Times, 11/12/2011)

The fully developed discourse lattice (see Figure 19) required to capture the conceptual nature of the

above example consists of three conceptual integration networks which constitute a metaphor system that incorporates two metaphors of conflict, and one sport metaphor: POLITICAL ARGUMENT IS A FIGHT, ELECTIONS ARE A STANDOFF, and ELECTIONS ARE A RACE. The first two integration networks have a shared generic space, i.e. their generic spaces are structured by the event structure metaphor, so that they consist of the following elements: setting, participants, means, and goal. The third network has a combined generic space39 that incorporates the previously mentioned elements of the event structure metaphor, and the SOURCE-PATH-GOAL image schematic structure. This image schematic structure is introduced by the activated entrenched compression of the RACE frame which is linked to the metaphor of JOURNEY and its underlying SOURCE-PATH-GOAL image schematic structure. In that sense, the image-schematic-based generic space is contained within the event-structure-metaphor-based generic space, thus explaining the notion of the combined generic space.

Each of the blended spaces inherits its organizing frame from the corresponding source input, rendering all three networks single-scope, with FIGHT, STANDOFF, and RACE as the organizing frames of the blends 1, 2, and 3, respectively. Additionally, in the discourse lattice, the attention shifts from Blend 1, to Blend 2, and finally to Blend 3, which means that the three blends are taking turns playing the role of the focus

space.40 Each blend develops emergent structure as a direct product of compression in the network. Namely,

the emergent structure is the result of Disanalogy between the means for achieving the goal in each of the pairs of inputs in the three networks, which is compressed into Uniqueness in each of the respective blends. Furthermore, emergent structures are constructed through the following stages in each of the networks:

39 The notion of the combined generic space will be further elaborated in the section dealing with SPORT metaphors. 40 See section 2.2.1 above for details.


Figur

e 19.

Metap

hor s

ystem

: FIG

HT, S

TAND

OFF,

and R

ACE


1. composition: in all three networks counterpart connections are established between the elements from the inputs, and these are then projected to the blend. 2. completion: in all of the networks structure from the target input is projected into the already pre-compressed topology of the corresponding source input (Fauconnier & Turner, 2002), and background knowledge structures related to the organizing frame of the source input are activated in order to facilitate this process. 3. elaboration: mental simulation of all three blends can be interpreted as coupled elaboration.

In general, the emergent structures in all three blended spaces can cause backward projections to the inputs and therefore increase the rhetorical force of each metaphorical expression. In the above presented discourse lattice structure of the discussed metaphor system (Figure 19), such backward projections can conspire to create an even greater amplification of the rhetorical effect. This means that the metaphor system licenses the interaction not only at the level of entrenched conceptualizations, but also in the process of active on-line meaning construction. This also opens up possibilities for the interaction of backward projections that need not remain confined to their original network, but may be transferred further through the discourse lattice. This can especially be the case for the first two networks, which are linked by the same conceptual key of CONFLICT. It may be the case that metaphorical integration networks linked by the same entrenched conceptual mechanism, i.e. the same overarching conceptual key, can manifest stronger interaction, and even serve as resonators to each other’s rhetorical force.

In that sense, interaction between the networks can actually forge the vital relation of Time between the three blends, which is introduced owing to the temporal and chronological organization imposed by the discourse structure, i.e. the immediate context of the above example. In that sense, the change of focus between the blends follows the discourse chronology, and the vital relation of Time actually facilitates this process and enables the readers to navigate their way through the discourse structure, and perform both compression and decompression as needed. Again, it is worth noting that all of the events presented in Figure 19 are taking place within the Republican Party during the preliminary elections that involve various interactions between the members of that party. In other words, both the conflicts and the race involve members of the same political group.

The Governing Principles of Compression that have been satisfied in all three networks are the same as in other metaphorical networks discussed so far, and they include the following: compression of one vital relation into another, borrowing for compression, and scalability. There is an additional vital relation of Time that is not compressed, but serves to provide a chronological link between the three blends. The systematic nature of compression can also be explored relying on the “molecule-model.” which would be extremely elaborate in this case, since there is an interaction between three conceptual integration networks.


(7) Senate Democrats introduced legislation Monday to extend and expand an expiring payroll tax cut, setting the stage for a showdown with Republicans who are almost certain to reject the Democrats’ proposal for paying for the cut. (The New York Times, 29/11/2011)

The metaphorical expression from the previous example illustrates the conceptual metaphor POLITICAL

NEGOTIATIONS ARE A SHOWDOWN, and the conceptual key POLITICS IS CONFLICT. The generic space is structured by the event structure metaphor as can be seen in Figure 20, while the blend adopts the organizing frame of the source input, making this another instantiation of a single-scope network. Furthermore, the blend develops emergent structure through the following processes: 1. composition: counterpart connections between the elements from the two inputs are established, and elements are projected to the blend. 2. completion: this process involves the recruitment of certain entrenched background knowledge structures related to the organizing frame of the blend, which aides in the completion of the structure of the blended space. 3. elaboration: bearing in mind that the elaboration of the present blend can affect real-time inferences and actions, this is another example of coupled elaboration.

Figure 20. Showdown between Democrats and Republicans

Additionally, the emergent structure appears as a direct product of compression of the outer-space vital relation of Disanalogy between the means for achieving the goals in the two input spaces. Such emergent


structure can also cause backward projections to the inputs, thereby increasing the rhetorical strength of the metaphor.

The most dominant vital relation is the outer-space vital relation of Analogy that is compressed into the inner-space vital relation of Uniqueness in the blended space. In other words, the vital relation of Analogy that connects the causal structures of the two input spaces is also compressed into Uniqueness in the blend, as is the vital relation of Intentionality. In terms of the Governing Principles of Compression, the present network satisfies the following: compression of one vital relation into another, borrowing for compression, and scalability.

Closely related to the notion of borrowed compression is the systematic nature of compression (in the sense of Fauconnier & Turner, 2008). Namely, the diagram in Figure 20 captures the on-line aspect of the process of meaning construction. However, beside such sculpting, the network also includes the elements of cobbling, i.e. the entrenched sections that are recruited, and which actually fuel the on-line process of meaning construction. This can be illustrated by a “molecule-model” diagram which can be developed for the fully-fledged model of the present conceptual integration network. (8) “I would like Monti to be a bit more courageous on this front, about the labor market,” said Sergio Romano, a

columnist for Corriere della Sera and a former ambassador. (The New York Times, 16/12/2011)

The above metaphorical expression can be attributed to the conceptual metaphor POLITICAL

NEGOTIATIONS ARE A CONFRONTATION ON THE WAR FRONT, which further corresponds to the conceptual key POLITICS IS CONFLICT. The generic space contains elements organized by the event structure metaphor, and they include: setting, participants, means, and goal. The clash between the organizing frames of the two input spaces is resolved by the projection of the organizing frame of the source input to the blend, rendering the present network single-scope.

Owing to the compression of Disanalogy between the means for achieving the goals in the two input spaces, the blended space develops emergent structure presented in Figure 21. This also involves the following processes: 1. composition: counterpart connections between the elements from the two inputs are established, and elements are projected to the blend. 2. completion: entrenched background knowledge structures are activated in order to facilitate the completion of the blend. 3. elaboration: mental simulation of the blend can lead to actual action caused by the real-time inferences, which makes it another instance of coupled elaboration. Such emergent structure can facilitate backward projections to the input spaces, thereby augmenting the rhetorical power of the metaphor (e.g. by provoking a stronger emotional response).


Figure 21. War front

The satisfied Governing Principles of Compression include the following: compression of one vital

relation into another, borrowing for compression, and scalability. Again, the most dominant is the compression of the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness in the blend, represented by the horizontal lines between the inputs in Figure 21. Reflecting the discussion presented in Fauconnier and Turner (2008), WAR FRONT, and POLITICAL NEGOTIATIONS represent pre-built entrenched compressions available for recruitment. In that sense, the diagram from Figure 21 represents only the on-line aspect of the process of meaning construction, i.e. the construction of the present conceptual integration network. For the remainder of the fully-fledged network to be included in the diagram would require the construction of the more elaborate “molecule-model” (see figure 14 for details). 4.1.2.1.2. Systematicity of Emergent Structure in CONFLICT Metaphors

Similar analyses to those performed in the examples above were conducted over the entire corpus. Such further in-depth analyses of the entire corpus of CONFLICT metaphors yielded a very relevant finding. Namely, 88.11% of cases (i.e. 126 metaphorical expressions) showed a consistent form of emergent structure that can be abstracted in the formula A harms B. The remaining 11.89% of cases showed different forms of emergent structure. More specifically, 10.49% (i.e. 15 cases) also showed consistency in the form - A prevents B from being harmed, while the remaining 1.4% of cases showed two different additional forms. The distribution of emergent structures is presented in Table 2.


Form of emergent structure Number of occurrences in the corpus Percentage

A harms B 126 88.11% A prevents B from being harmed 15 10.49%

Other 2 1.40%

Table 2. Emergent structures in CONFLICT metaphors

The established systematic nature of the emergent structure in CONFLICT metaphors can be attributed to the highly entrenched nature of the conceptual key POLITICS IS CONFLICT that binds them together. Namely, the entrenched mechanism of the conceptual key conditions the choice of the source and target inputs of conceptual integration networks and can be argued to facilitate the alignment of topologies of the inputs. In line with the notion of borrowed compressions, it could be suggested that the entire networks and the structures of the blended spaces are actually governed by the borrowed conceptualization of the conceptual key, where the particular entrenched conceptualization can itself be understood as a pre-built compression available for recruitment. In that sense, the structures of the blends, including their emergent structures, should in fact show certain systematicity, owing to the entrenched, i.e. pre-compressed nature of the governing conceptual key they represent. Furthermore, such systematicity in emergent structures also contributes to metaphor coherence in metaphor systems constructed out of CONFLICT metaphors, and facilitates their mutual interaction (see examples 19, 23, and 24).

In the sense of Fauconnier and Turner (2008), the systematic nature of compression seems to lead to the systematicity in the form of emergent structures. Namely, the background knowledge structures that fuel the on-line process of meaning construction are conditioned by the context introduced by the entrenched conceptualization of the conceptual key, and are in that sense semantically related. Such relatedness further influences the active computations performed on-line, and conditions the systematicity of the resulting structures.

The two abstract formulaic representations of emergent structure presented in Table 2 can be understood as recursive patterns. However, it needs to be immediately stressed that it is not recursion in the generativist sense of pre-defined abstract grammatical structural patterns that await to be filled with proper elements. Instead, this is the case of recursion at the conceptual level. Namely, the two abstract formulas from Table 2 make up a recursive conceptual mechanism that can be used to describe the semantic content of the emergent structures, and not their grammatical structural patterns. In that sense, these formulas are somewhat similar to the general patterns represented by conceptual metaphors and conceptual keys (e.g. ELECTIONS ARE A BATTLE), suggesting that the relationship between the formula A harms B, and the corresponding emergent structures is the same as the relationship between a conceptual metaphor, and its corresponding metaphorical expressions. Such reasoning sanctions the conclusion that conventional, i.e. entrenched, conceptual metaphors and conceptual keys can also be treated as recursive cognitive


Metaphorical expression Emergent structure Form of emergent structure

(9) After months of relative quiet, the Republican presidential candidates have started to increase their television advertising, bringing a new competitive dynamic to a fight that has largely remained off the commercial airwaves until now. ( 6/12/2011)

A good TV election campaign harms political rivals

A harms B

(10) In the battle over how to save the euro, Germany has plenty of leverage but not many friends. 9/12/2011

Political negotiations can harm participants

A harms B

(11) On the political front, Mr. Wang began a campaign for “Happy Guangdong,” derided by critics as empty sloganeering. (31/12/2011)

Political campaign harms political rivals

A harms B

(12) Normally, a candidate spends the dwindling hours before the Christmas break in states that vote first, like Iowa, where the voting starts in less than two weeks and where Mr. Gingrich is dropping in the polls under a barrage of negative attacks. (20/12/2011)

Negative campaign harms Gingrich

A harms B

(13) The standoff pits Prime Minister Nuri Kamal al-Maliki, a Shiite, against one of his most nettlesome partners in Iraq’s government, the Iraqiya coalition, a multisectarian group with wide support among secular Iraqis and Sunni Muslims. (18/12/2011)

Political fractions harm each other with arguments (i.e. political arguments harm political fractions)

A harms B

(14) Mr. Romney had spent the week leading up to the debate attacking Mr. Gingrich — in interviews, television and Web ads, and through surrogates — but during the debate on Thursday, he tamped down his criticisms. (17/12/2011)

Romney’s arguments harm Gingrich (i.e. political arguments harm the participants of the political process)

A harms B

(15) A political showdown now looms over the Lokpal issue, possibly determining whether the Congress Party can reverse its sagging political fortunes midway through the government’s five-year term. (23/12/2011)

Arguments harm political rivals (i.e. argument over the Lokpal issue can harm political rivals)

A harms B

(16) Mrs. Merkel was “fighting for the future of Europe, and we stand behind her,” Mr. Brüderle said Friday. (3/12/2011)

Political arguments harm political rivals

A harms B

(17) Turkey has been locked in an intractable political fight with Cyprus since 1974, when it invaded the island to prevent a proposed union with Greece and set up a rival government in the ethnic Turkish part of Cyprus that only it recognizes. (5/12/2011)

Political arguments harm the participants of the political process

A harms B

(18) While treaty change can be a lengthy process, the hope is that the effort will create enough momentum for economic convergence and discipline that will provide the political cover for Germany’s leaders to allow the European Central Bank to step in much more forcibly to defend Italy and Spain and try to stabilize the market. (2/12/2011)

ECB protects EU member states (Italy and Spain) from being harmed

A prevents B from being harmed

(19) Bowing under intense pressure from members of their own party, House Republican leaders agreed Thursday to accept a temporary extension of the payroll tax cut, beating a hasty retreat from a showdown that Republicans increasingly saw as a threat to their election opportunities next year. (23/12/2011)

Accepting the extension of the payroll tax cut prevents the Republicans from being harmed (i.e. good political strategy prevents the Republicans from being harmed)


(20) But as Mr. Gingrich tours the state by bus in the final days of an increasingly bitter campaign, he is wielding Nice Newt — or trying to — as a kind of last defense against a shelling of attack ads, ridicule from opponents and a drip-drip-drip of tough news coverage that has clearly blunted his surge to near-front-runner status. (29/12/2011)

Gingrich resorts to campaign strategy to prevent himself from being harmed


Table 3. Emergent structures in CONFLICT metaphors: Additional examples


mechanisms. This property can serve to further account for the “diagnosed” systematicity in the form of emergent structure in the present corpus.

Additionally, Coulson and Oakley (2005) suggested that with novel metaphors emergent structures are completely computed on-line, whereas in conventional metaphors, emergent structures are recruited. In that sense, and using the terminology of Fauconnier and Turner (2008), it can be said that the previously described semantic-conceptual recursive mechanisms (e.g. A harms B) constitute the entrenched part of conventional metaphors which actually represent a pre-built, i.e. pre-compressed, conceptual mechanisms available for recruitment. These pre-built compressions then acquire their elements and specific structure based on the context imposed by the organizing frame of the blend, and the compressions of elements from input spaces. In other words, the conceptual recursive pattern (e.g. A harms B) is completed via the process of pattern

completion, which, in addition to compressions of vital relations, constitutes a central process in the formation of emergent structures in conventional metaphors.

Some additional examples of metaphorical expressions from the corpus along with their emergent structures are presented in Table 3. 4.1.2.1.3. Function and Emotional Appeal of CONFLICT Metaphors

Political discourse in general is highly marked by its persuasive dimension. One of the salient rhetorical tools used to facilitate the process of persuasion is the use of conceptual metaphors (Lakoff, 1991; Mio, 1996; 1997; 2005; Thompson, 1996; Charteris-Black, 2004; 2009; 2011; Burnes, 2011; Bougher, 2012). As argued in section 2.1.10, the main reason for this resides in the fact that metaphors can be used to highlight or hide only the desired aspects of an issue (in the sense of Lakoff & Johnson, 2003[1980b]). Additionally, metaphors can be used to simplify complex issues and render them understandable to a wider population, or they can provoke an emotional reaction, which makes the public more susceptible to manipulation (e.g. Mio, 1997; Charteris-Black, 2009; 2011; Bougher, 2012).

In the case of CONFLICT metaphors, one of the dominant mechanisms that license their persuasive function is the ability to provoke an emotional response. Namely, “the war metaphor is ubiquitous, connected to strong emotions and social values and it is widely used in politics of mass appeal” (Steinert, 2003: 268), where “emotion is a crucial feature to be taken into consideration as well with regard to the strategic use of metaphor” (Ferrari, 2007: 611). Furthermore, “population and politicians share common value orientations such as ‘(patriarchal) family/community’ and ‘warrior/masculinity’” (Steinert, 2003: 267), which is why politicians often aim to exploit those values for personal benefit.

By placing the political process in the frame of CONFLICT, it is possible to talk about political negotiations as fights, attacks, battles, standoffs, showdowns, defense, or (war) fronts (examples 9 – 20 above). Such framing of political issues “creates pressure for unity, solidarity, mobilization of people and resources for the common good (against the foe)” (Steinert, 2003: 268). In other words, the CONFLICT


metaphor has the power to bring people together in a collective effort of self-preservation. Additionally, bearing in mind that “war forges the bond of community and acceptance of (political) leadership like nothing else” (Steinert, 2003: 266), metaphors of CONFLICT can create a framework in which the public can be manipulated as a collective mass. Furthermore, these metaphors can also make the public empathize with their political favorites, which is evident in examples 12 and 14, where Mr. Gingrich is presented as the victim of attacks.

In the context of the CBT, the ability of metaphors to provoke an emotional reaction can be attributed to backward projections from the blend back to the inputs. Furthermore, the coupled elaboration of the emergent structure can produce real-time inferences that can further affect behavior and reasoning in relation to certain issues. For instance, example 10 presents the efforts made by Germany to protect the EU, in terms of a battle, which can result in the actual emotional involvement of the readers, since the notion of “battle” stresses the seriousness of the crisis, and simplifies the complex political and economical mechanisms that caused it in the first place by representing them in terms of an enemy. In turn, this process reflects Harder’s (2005) notion of polarization.41 In that sense, such emotional involvement can actually amplify the rhetorical power of the metaphor. 4.1.2.1.4. Metaphor Systems and Metaphor Interaction

As discussed above, apart from instances of individual metaphors, there are also metaphor systems (in the sense of Lakoff, 1991; 1993; Lakoff & Johnson, 2003[1980]; Evans & Green, 2006; Kövecses, 2010). Percentage-wise, 29.37% of CONFLICT metaphors from the present corpus appear in elaborate metaphor systems. Apart from reflecting the notions of metaphorical coherence42 from the CMT framework, individual metaphors that comprise the metaphor systems can serve as resonators to each other’s rhetorical force, which results in an amplified rhetorical effect. In addition to backward projections, section 4.1.2.1.1 has shown that immediate discourse context can actually serve to forge vital relations between the blended spaces of the metaphorical conceptual integration networks comprising the system. That way, apart from serving as focus spaces in the discourse lattice, these metaphorical blends seem to procure a deeper link. Additionally, such connections between the metaphorical blends can be argued to facilitate the process of achieving human

scale at the discourse level.

(21) But they are hoping to help stretch the Republican nominating contest into a longer and bloodier battle — meaning they are eager to define Mr. Gingrich for voters in unflattering terms without necessarily wounding him fatally and assisting Mr. Romney, whom they still view as a formidable general election opponent. (The New York Times, 8/12/2011)

(22) Mindful of some Republicans’ goal of privatizing Social Security, the Democrats suggested that tinkering with the payroll tax “may be used as the first step in a larger battle to fundamentally dismantle Social Security.” (The New York Times, 16/12/2011)

41 See section 3.2 above for details. 42 See section 2.1.4 above for details.


(23) Mr. Compagnone is one soldier in a battle - often uphill - to persuade Italy's famously tax-evading citizens to pay up. (The New York Times, 25/12/2011)

(24) It is unclear what concrete changes will flow from the decision. A top Communist Party official, Ivan Melnikov, told the newspaper Vzglyad that Mr. Volodin’s selection signaled a shift in the direction of “a more harsh model for the coming political battle.” (The New York Times, 28/12/2011)

(25) They have controlled the spending conversation from the first battle, when new members demanded that Republicans hold out for billions more in cuts than those sought by House appropriators. It continued through the debt ceiling showdown, in which a plan to cut $2.1 trillion over 10 years was formed, as well as during the intense focus on changes in entitlement programs. (The New York Times, 4/12/2011)

(26) “London is the center of financial services in Europe,” Mr. Cameron said in October. “It’s under constant attack through Brussels directives. It’s an area of concern; it’s a key national interest that we need to defend.” (The New York Times, 10/12/2011)

(27) But European officials believe that a significant move toward treaty changes to create more economic governance in the euro zone — with tighter, more enforceable limits on debt and centralized oversight of national budgets — will give the European Central Bank the political cover to act more aggressively to defend Italy and Spain and drive down currently unsustainable interest rates on their bonds. (The New York Times, 2/12/2011)

(28) Its shifting alliances, reflecting different currents within the movement, helped keep Islah ahead of its opposition rivals in Yemen. (The New York Times, 3/12/2011)

(29) And while even other Republicans were predicting that the House Republicans would have to blink, or risk further political damage, the ugliness of the fight reminded Americans yet again of the seeming futility of Mr. Obama’s 2008 campaign promise to make Washington work as the year of his re-election race is upon him. (The New York Times, 22/12/2011)

(30) But in an acknowledgement that he might not be able to reverse Mr. Gingrich’s momentum quickly, Mr. Romney and his team are bracing for a far rougher slog through the early Republican nominating contests than they had envisioned even a few weeks ago and preparing for months of a state-by-state, delegate-by-delegate fight. (The New York Times, 15/12/2011)

Previous examples illustrate the interaction between various conceptual metaphors, some of which belong to different conceptual keys. For instance, examples 21, 25, and 26 represent interactions between different CONFLICT metaphors; examples 22, 23, 24, and 28 represent interactions between CONFLICT and JOURNEY metaphors; example 29 represents interaction between CONFLICT and SPORT metaphors; example 30 represents interaction between JOURNEY, SPORT, and CONFLICT metaphors; example 27 represents interaction between JOURNEY, CONTAINER, and CONFLICT metaphors.

SPORT Metaphors 91

4.1.2.2. SPORT Metaphors SPORT metaphors will be organized under the conceptual key POLITICS IS SPORT, and they take up

40.42% of the corpus, with 97 metaphorical expressions. This conceptual key will include conceptual metaphors and their corresponding metaphorical expressions organized around the following metaphor keywords: race, team, jab, players, sparring, playing field, game, etc. A complete overview of metaphor keywords corresponding to the conceptual key POLITICS IS SPORT and their distribution in the present corpus are presented in the bar graph in Figure 22. The most dominant metaphor keyword under the conceptual key of SPORT was ‘race’, which reflects the highly competitive nature of primary elections that were under way in the US in December 2011. Additionally, the recorded frequent use of SPORT metaphors in the present corpus is in line with the results from previous research in the field (e.g. Howe, 1988; Semino & Masci, 1996; Silaški & Radić-Bojanić, 2010; Radić-Bojanić & Silaški, 2008).

Figure 22. Distribution of metaphor keywords in SPORT metaphors

4.1.2.2.1. Structure of SPORT Metaphors The present section of the paper will discuss the structure of SPORT metaphors by addressing some

specific examples from the corpus, in accordance with the general guidelines for the analysis described in section 4.1.1.

SPORT Metaphors 92

(31) By virtually all accounts, the 2012 presidential race was to hinge on a restless electorate’s overriding worry, the troubled American economy. (The New York Times, 30/12/2011)

The present example is an instance of a conceptual metaphor PRESIDENTIAL ELECTIONS ARE A RACE, which can be further attributed to the conceptual key POLITICS IS SPORT. The conceptual integration network presented in Figure 23 has a combined generic space, where the generic space contains elements that include setting, participants, goal, and means, which originate from the event structure metaphor, as well as the SOURCE-PATH-GOAL image schematic structure that reflects the image schematic nature of the RACE metaphor. In line with the discussion presented in section 2.2.3, such generic space organization is aimed at foregrounding the role of conceptual primitives (in the sense of Mandler, 1992), i.e. the role of image schemas in the construction of the human cognitive architecture that supports the previously introduced entrenched conceptual pattern PRESIDENTIAL ELECTIONS ARE A RACE.

Figure 23. Presidential race

The source input is organized by the RACE frame, while the target input is organized by the frame of PRESIDENTIAL ELECTIONS. It is worth noting that the presidential candidates in this case belong to two opposing political parties, the Republicans and the Democrats. The blend inherits the organizing frame of the source input, rendering the present network single-scope, which means that the elements from the target input are projected into the frame of the source input. The blend also develops its central emergent structure through the following processes: 1. composition: the network establishes counterpart connections between the elements from the two inputs, and these are projected into the blend;

SPORT Metaphors 93

2. completion: the blend recruits background knowledge structures related to the organizing frame of the source input, i.e. the organizing frame of the blend, which completes its topology; 3. elaboration: again, this is the case of coupled elaboration, since it affects real-time inferences, reasoning, and behavior.

The developed emergent structure can give way to backward projections to the inputs, which can serve to amplify the initial rhetorical force of the metaphor. In terms of compressions, the emergent structure can be understood as the direct product of compression of the outer-space vital relation of Disanalogy between the means for achieving the goals in the two respective inputs, into the inner-space vital relation of Uniqueness in the blend. Additionally, the causal and intentional structures, as well as the elements from the two inputs are connected by the vital relation of Analogy that is compressed into Uniqueness in the blend. When it comes to the Governing Principles for Compression, the following are satisfied: compression of one vital relation into another, borrowing for compression, and scalability.

In terms of the systematic nature of compression (in the sense of Fauconnier and Turner, 2008), both inputs can be understood as prebuilt compressions, i.e. as results of previous conceptual integration work, available for recruitment. In that sense, the network presented in Figure 23 reflects only the sculpting process, whereas a fully developed “molecule-model” similar to the one in Figure 14 would be required to also capture the cobbling aspect of the present network. (32) Only as the first voting draws closer is the pace of traditional campaigning picking up, with candidates trying

to make up for lost time. (The New York Times, 1/12/2011)

The above example can be attributed to the conceptual metaphor PRIMARY ELECTIONS ARE A RACE, which further belongs to the conceptual key POLITICS IS SPORT. The combined generic space is structured like in the previous example, and such generic space structure is typical of all RACE metaphors, which will be evident from the forthcoming analyses.

The conceptual clash at the frame-level is resolved by projecting the organizing frame of the source input as the organizing frame of the blend, which makes this network single-scope. Elements from the target input are therefore projected into the pre-compressed organizing frame of the source input. The blend also develops its emergent structure though the following processes: 1. composition: counterpart connections are established between the elements from the inputs, and these are projected, i.e. compressed in the blend; 2. completion: the blend recruits background knowledge related to the RACE frame, which serves to complete the blend; 3. elaboration: this is another instance of coupled elaboration, since the emergent structure can affect reasoning and behavior in real-time. Additionally, such emergent structure can yield backward projections to the inputs, which can serve to provoke an emotional response and therefore amplify the rhetorical force of the metaphor. In particular, the RACE

SPORT Metaphors 94

metaphor reduces the complexity of a political debate to a simple contest, which in turn involves supporters on an emotional level.

Figure 24. The pace of traditional campaigning is picking up

The present emergent structure can also be understood as a direct result of the compression of the

outer-space vital relation of Disanalogy between the means for achieving the goals in the two inputs, into the inner-space vital relation of Uniqueness in the blend. As was the case in other metaphorical networks, the most dominant compression is that of the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness. Compression of one vital relation into another, scalability, and borrowing for compression have all been satisfied.

Directly related to the principle of borrowing for compression is the systematic nature of compression. Namely, both inputs are instances of pre-built compressions available for recruitment, and the network depicted in Figure 24 reflects only the active on-line process of conceptual integration. Like in previous examples, a more elaborate “molecule-model” network structure would reveal the ‘hidden’ cobbling aspects of the present network. (33) The holiday brinkmanship over the issue recalled the December budget showdown 16 years ago between

another first-term Democratic president, Bill Clinton, and a new Republican Congressional majority — a fight that capped their year of confrontation over the nation’s fiscal priorities by reviving Mr. Clinton politically as he began his re-election race. (The New York Times, 22/12/2011)

The present example shows interaction between three conceptual metaphors: POLITICAL ARGUMENT IS A

SHOWDOWN, POLITICAL ARGUMENT IS A FIGHT, and RE-ELECTION CAMPAIGN IS A RACE. The first two conceptual metaphors can be attributed to the conceptual key POLITICS IS CONFLICT, whereas the third one can be viewed

SPORT Metaphors 95

Figur

e 25.

Clint

on’s

re-e

lectio

n rac

e

SPORT Metaphors 96

as an instance of the conceptual key POLITICS IS SPORT. The generic spaces of all three networks are typical instances of generic spaces found in CONFLICT and RACE metaphors, respectively, both of which have already been discussed in the previous analyses. Additionally, all three networks from Figure 25 are instances of single-scope networks, suggesting that the organizing frames of their respective blends are adopted from the corresponding source inputs, so that the elements from the target inputs are projected into the pre-compressed topologies of the organizing frames of their corresponding source inputs. Additionally, each blend develops its emergent structure through the processes of composition, completion, and elaboration, in accordance with the general principles discussed in the previous examples. Each of the emergent structures provides a possibility for backward projections to its respective inputs, which can in turn serve to augment the rhetorical power of the given metaphor. Furthermore, the immediate discourse context forges an additional ‘bond’ between the three metaphors that comprise the present system, thereby allowing them to serve as resonators to each other’s rhetorical force. In other words, the backward projections need not be restrained to their ‘mother network’, but can also work across networks.

The three blends in the network presented in Figure 25 make up the fully developed discourse lattice, and are connected by the change of focus, where the discourse ‘moves’ from Blend 1, to Blend 2, and finally to Blend 3. These changes of focus can be argued to provide an additional topological link in the entire network, and thus serve to further enhance the rhetorical force of the present metaphor system. Owing to the high degree of relatedness between the frames of showdown and fight, which serve as the organizing frames of blends 1 and 2, respectively, it can be argued that in addition to the simple change of focus, the immediate discourse context also forges a vital relation of Analogy between the first two blends. Namely, the frames of showdown and fight belong to the same conceptual key of CONFLICT, and both the topologies and causal structures that they introduce into the blends are analogous. Furthermore, as a result of these analogies, the emergent structures of blends 1 and 2 are identical, suggesting that the immediate discourse context also leads to the creation of the vital relation of Identity between the emergent structures of the first two blends. Such network structure also increases metaphor coherence, i.e. the coherence between the individual networks comprising the system.

Emergent structures in the three networks can be understood as direct products of compression, more precisely the compression of Disanalogy into Uniqueness, as was the case in all previous analyses. Principles of compression of one vital relation into another, scalability, and borrowing for compression have all been satisfied. The systematic nature of compression for the present metaphor system can be represented by a fully developed “molecule-model” that would capture both the on-line process of meaning construction depicted in Figure 25, and the process of background knowledge recruitment in the form of pre-built compressions.

SPORT Metaphors 97

(34) In the meantime, Mr. Paul’s rivals continued their last-minute sprints across Iowa in the hopes of rallying support among an electorate that appears to remain open to persuasion. (The New York Times, 31/12/2011)

This is an instance of a conceptual metaphor PRIMARY ELECTIONS ARE A RACE which belongs to the

conceptual key POLITICS IS SPORT. The combined generic space of the present network has all the characteristics already discussed in examples 31 and 32 above. The blend adopts the organizing frame of the source input, which makes the present network single-scope, with all the typical traits of single-scope networks discussed so far. Emergent structure is developed through the familiar processes of composition, completion, and elaboration, and this is another case of coupled elaboration. The emergent structure also licenses backward projections to the inputs, which can result in an amplified rhetorical force of the metaphor, since such backward projections can provoke an emotional appeal with the audience.

Figure 26. Last-minute sprints

The emergent structure of the present network can be dubbed a direct product of compression of the outer-space vital relation of Disanalogy between the means for achieving the goals in the inputs, into the inner-space vital relation of Uniqueness in the blend. The causal and intentional structures of the two inputs, as well as their elements, are connected by the vital relation of Analogy that is compressed into Uniqueness in the blend. In terms of the Governing Principles for Compression, the network satisfies compression of one vital relation into another, scalability, and borrowing for compression. The systematic nature of compression can be captured by a “molecule-model” that would reflect both the cobbling and sculpting aspects of the present network, owing to the fact that both inputs can be understood as instances of pre-built compressions.

SPORT Metaphors 98

(35) Former Gov. Jon M. Huntsman Jr. of Utah, seeking to break out of the Republican pack, appeared to jar the audience when, in talking about the lack of trusting government, he declared: “We have been kicked around as a people. We are getting screwed as Americans.” (The New York Times, 16/12/2011)

This is another instance of a conceptual metaphor PRIMARY ELECTIONS ARE A RACE that belongs to the

conceptual key POLITICS IS SPORT. The present network also has a combined generic space already discussed above. The blend adopts the organizing frame of the source input, rendering the present network single-scope. This suggests that elements from the PRIMARY ELECTIONS frame are projected into the pre-compressed topological structure of the RACE frame. The central emergent structure of the blend, which presents itself as a direct product of compression, is developed through the processes of composition, completion, and elaboration. As with other instances of conceptual metaphors in political discourse, this is again the case of coupled elaboration. Such emergent structure facilitates backward projections from the blend to the inputs, which can increase the rhetorical power of the metaphor by, for example, provoking an emotional response.

Figure 27. Break out of the pack

Emergent structure can also be viewed as the direct product of compression of the outer-space vital

relation of Disanalogy between the means for achieving the goals in the inputs, into the inner-space vital relation of Uniqueness in the blend. Again, the most dominant compression is that of the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness in the blend. Namely, causal and intentional structures, as well as the counterparts in the two inputs are analogous. Additionally, the two inputs are also obvious instances of pre-built compressions available for recruitment, and the network from Figure 27 presents only the on-line aspect of meaning construction. Activation of background knowledge structures and

SPORT Metaphors 99

the massive work on conceptual integration that looms in the background can be presented by a more elaborate “molecule-model,” similar to the one in Figure 14.

(36) ''When a principled conservative took the lead, they outspent Newt Gingrich 20 to 1, attacking him with

falsehoods.'' (The New York Times, 28/12/2011) The above example shows an interaction between two conceptual metaphors: PRIMARY ELECTIONS ARE

A RACE, and PRIMARY ELECTIONS ARE A BATTLE, where the former can the attributed to the conceptual key POLITICS IS SPORT, and the latter to the conceptual key POLITICS IS CONFLICT. The first network has a combined generic space typical of RACE metaphors, whereas the second network has the generic space based on the event structure metaphor, typical of CONFLICT metaphors. Both networks are single-scope, meaning that the organizing frames of the two blends are adopted from their respective source input spaces. Each of the blends develops its own emergent structure through processes of composition, completion, and elaboration. Additionally, both blends are instances of coupled elaboration, where the developed emergent structure can influence real-time reasoning and behavior. The emergent structures can also give way to backward projections that can serve to amplify the rhetorical forces of the respective metaphors. Furthermore, bearing in mind that this is another instance of a metaphor system, the two blends can serve as resonators to each other, thus increasing each other’s rhetorical power, which is licensed by both the immediate discourse context, and the coherence of the two metaphors.

The emergent structure in both blends can be presented as a direct consequence of compression of the outer-space vital relation of Disanalogy between the means for achieving the goals in the two networks, respectively, into the inner-space vital relation of Uniqueness in the two corresponding blends. Again, both networks are dominated by the compression of the outer-space vital relation of Analogy that exists between the causal and intentional structures of the inputs, and the counterpart elements of the inputs, into the inner-space vital relation of Uniqueness in the two respective blends. Additionally, both networks satisfy the following principles of compression: compression of one vital relation into another, scalability, and borrowing for compression. Furthermore, the systematic nature of compression could also be depicted by an elaborate “molecule-model” which would include both the on-line sculpting of the network and the background cobbling that facilitates the process of conceptual integration.

In addition to the change of focus between the two blends in the discourse lattice, an additional vital relation of Cause-Effect is also formed between them as the direct result of the immediate discourse context. Namely, the attacks on Gingrich result directly from his perceived superior position, which consequently leads to a more profound link between the two blends. It can be argued that the additional vital relation between the two blends amplifies not only their rhetorical force and coherence, but increases the discourse coherence as well, and facilitates the process of meaning construction.

SPORT Metaphors 100

Figur

e 28.

Ging

rich t

ook t

he le

ad

SPORT Metaphors 101

(37) Mrs. Merkel, a central player in efforts to rescue Europe’s single currency, was addressing the German Parliament before a meeting next week in Brussels, when Europe’s leaders will try again to find a politically palatable solution to the crisis. (The New York Times, 3/12/2011)

The present example is an instance of interaction between a SPORT metaphor and a CONFLICT

metaphor. Both networks are instances of single-scope networks, where the organizing frames of the two blends are acquired from their respective source inputs. Additionally, blends 1 and 2 share the same generic space based on the event structure metaphor, which consists of the following elements: setting, participants, goal, and means.

Both blends develop their emergent structures through the processes of composition, completion, and elaboration. Additionally, the two blends show traits of coupled elaboration, where the elaboration of the blend influences reasoning and behavior in real-time. These emergent structures can also facilitate backward projections from the blends to their respective inputs. More specifically, blends 1 and 2 develop their emergent structures owing to the outer space vital relation of Disanalogy between the means for achieving the respective goals, which is compressed into Uniqueness in each of the two blends. The most dominant compression in the two networks is that of Analogy into Uniqueness, as was the case in all other metaphorical networks discussed so far.

Apart from being connected by the change of focus in the discourse lattice, the immediate discourse context seems to forge a more profound link between the two blended spaces. Namely, bearing in mind their topologies, and the causal and intentional structures that govern those topologies are governed by, it can be argued that the immediate discourse context forges an additional outer space vital relation of Analogy that exists between blends 1 and 2. Consequently, their emergent structures are also analogous to each other. Owing to this additional outer space vital relation of Analogy, backward projections need not be restrained to their ‘original’ networks. Instead, the two blends can serve as resonators to each other’s rhetorical power, which can, for instance, serve to amplify the audience’s emotional reaction, or to fuel other mechanisms of persuasion.

As was the case with all examples of metaphorical conceptual integration networks discussed so far, both networks in the present metaphor system are dominated by the outer space vital relations of Analogy, and their compressions into the inner-space vital relation of Uniqueness in each of the blends, respectively. Additionally, the two networks also satisfy the principles of compression of one vital relation into another, scalability, and borrowing for compression. Bearing in mind the systematic nature of compression (in the sense of Fauconnier and Turner, 2008), the present network from Figure 29 could be elaborated even further, in order to account for both the on-line process of meaning construction, and for the recruitment of the pre-existent compressions in the form of background knowledge structures. Such an approach would yield a diagram similar to that in Figure 14.

SPORT Metaphors 102

Figur

e 29.

Ange

la Me

rkel a

s the

centr

al pla

yer

SPORT Metaphors 103

(38) In their sparring Monday, Mr. Romney threw the first jab, sending an e-mail to supporters in which he called his rival an “unreliable leader,” who had supported action on climate change. (The New York Times, 13/12/2011)

The above metaphorical expression corresponds to the conceptual metaphor PRIMARY ELECTIONS ARE A

BOXING MATCH, which further corresponds to the conceptual key POLITICS IS SPORT. Although notions of ‘sparring’ and ‘jab’ are essentially related to the frame of conflict, a distinction is to be made between the notion of conflict within the BOXING frame, and the notion of conflict within the CONFLICT frame, since boxing involves a ‘controlled’ confrontation that takes place in accordance with a strict set of rules. In this particular case, it is important to stress the fact that the match is taking place between two members of the same political party.

Figure 30. Sparring between Romney and Gingrich

The generic space is organized by the event structure metaphor, similar to the generic spaces

discussed in CONFLICT metaphors. The organizing frame of the blend is adopted from the source input, suggesting that this is another instance of a single-scope network, where the elements from the target input are projected to the pre-compressed topology of the source input. The emergent structure of the blend is developed through processes of composition, completion, and elaboration, and this is another instance of coupled elaboration. Such emergent structure also offers the possibility of backward projections that can serve to amplify the rhetorical force of the metaphor.

There is a vital relation of Analogy that exists between the counterpart elements from the two inputs, which is compressed into Uniqueness in the final blend. Additionally, analogy can also be established between

SPORT Metaphors 104

the causal and intentional structures of the two input spaces. Compression of the vital relation of Disanalogy that exists between the means for achieving the goals in the two respective inputs into Uniqueness gives way to the development of emergent structure. It is also worth noting that the emergent structure from the present network is the same as the emergent structures developed in the majority of CONFLICT metaphors, i.e. it can be abstracted in the form A harms B. This can be explained by the already mentioned link between the frames of CONFLICT and BOXING.

The present network satisfies the following Governing Principles for Compression: compression of one vital relation into another, scalability, and borrowing for compression. Directly related to the notion of borrowing for compression is the systematic nature of compression, where a “molecule-model” could be used to account for both the on-line process of meaning construction, and for the recruitment of pre-built compressions.

(39) In the end, the prime minister [Mariano Rajoy] opted for a team containing several party veterans, including

some, like Mr. de Guindos, who already formed part of the government of José María Aznar, Spain’s last conservative prime minister. (The New York Times, 22/12/2011)

Figure 31. Team

The present example is an instance of a conceptual metaphor GOVERNMENT IS A TEAM, which further corresponds to the conceptual key POLITICS IS SPORT. The generic space is organized in a similar fashion to the generic spaces in CONFLICT metaphors discussed above, i.e. by the event structure metaphor. The organizing frame of the blend is that of TEAM, suggesting that this is an instance of a single-scope network. The emergent structure of the blend is developed through the processes of composition, completion, and elaboration, and the present network is another example of coupled elaboration.

SPORT Metaphors 105

In terms of compression, the present network develops its emergent structure owing to the compression of the vital relation of Disanalogy between the means for achieving the goals in the two inputs, into Uniqueness in the final blend. Causal and intentional structures, and the counterpart elements from the input spaces are connected by Analogy, and the most dominant compression in the present network, as was the case in all other metaphorical networks discussed so far, is the compression of the outer-space vital relation of Analogy into the inner-space vital relation of Uniqueness.

The Governing Principles for Compression satisfied by this network include compression of one vital relation into another, scalability, and borrowing for compression. In light of the systematic nature of compression discussed in Fauconnier and Turner (2008), the diagram in Figure 32 shows only the sculpting aspect, i.e. the live, on-line process of meaning construction. In order for the entire process to be captured, a “molecule-model” would need to be developed, which would encompass both the on-line aspects of processing, and the introduction of pre-built compressions via the process of schema induction (in the sense of Evans and Green, 2006).

4.1.2.2.2. Systematicity of Emergent Structure in SPORT Metaphors A detailed analysis of the remaining SPORT metaphors from the corpus revealed a certain degree of

systematicity of emergent structures. Unlike CONFLICT metaphors where the two most dominant forms of emergent structures were closely related, SPORT metaphors showed a greater number of source inputs structured by different frames, which in turn led to a greater diversity in the form of emergent structures, and these results are presented in Table 4.

Form of emergent structure Number of occurrences in the corpus Percentage

A improves the position of B 67 69.07% A harms B 11 11.34%

A and B are united by a common goal 9 9.28%

Other 10 10.31%

Table 4. Emergent structures in SPORT metaphors The most frequent form of emergent structure A improves the position of B was predominantly

associated with RACE metaphors, but was also developed in the context of some other metaphor keywords (see examples 37, 44, 45, 46, and 47). Such emergent structure is highly marked by its spatial topology, owing to its transparent links to the SOURCE-PATH-GOAL image schema that makes up an integral part of the above-discussed combined generic spaces. This connection is licensed by the borrowed compressions from the RACE frame which also has a marked spatial topology, inherently linked to the spatial image schematic structures. Consequently, this evidence supports the earlier claims of image schemas being an integral part of the human

SPORT Metaphors 106

Metaphorical expression Emergent structure Form of emergent structure

(40) A presidential race that has seen candidates abruptly rise and sharply fall is still remarkably unsettled here in Iowa, where the Republican nominating contest opens on Jan. 3. (7/12/2011)

Good campaign and arguments improve candidates’ positions in the primary elections

A improves the position of B

(41) After a dozen debates in the Republican primary race, which ranged from cantankerous to contentious, a forced sense of restraint hung in the air throughout the two-hour debate at the Sioux City Convention Center. (16/12/2011)

Good campaign and strong arguments improve candidates’ positions in the primary elections


(42) Mr. Gingrich is rated more favorably than any of the other six remaining candidates in the race among voters who say they are likely to attend the Republican caucuses in Iowa.7/12

Good campaign and the electorate’s support improve Gingrich’s position in the primary elections


(43) “Therein lies the quandary: With the caucuses right around the corner, do you want to let up on the gas pedal?” said Tim Albrecht, a spokesman for Gov. Terry E. Branstad of Iowa, a Republican. (23/12/2011)

Good campaign improves candidates’ positions in the primary elections


(44) The biggest player by far in the state has been Restore Our Future, a “super PAC” supporting Mr. Romney. (20/12/2011)

Strong support in the campaign improves Romney’s position in the primary elections


(45) While Mr. Biden runs the administration’s policy on Iraq, he does not have as central a role on Iran. But Mr. Biden, officials said, has been an influential voice in dealing with the upheaval in the Arab world, because he has dealt with many of the players as chairman of the Senate Foreign Relations Committee. (3/12/2011)

Good arguments and his background in the Arab politics improve Mr. Biden’s position in influencing the piece process in the Arab world


(46) Laying out a populist argument for his re-election next year, President Obama ventured into the conservative heartland on Tuesday to deliver his most pointed appeal yet for a strong governmental role through tax and regulation to level the economic playing field. (7/12/2011)

Persuasive arguments and a good campaign improve Obama’s position in the re-election process


(47) [Rick Santorum]“Our phone calls are going better, our volunteers are showing up, the crowds are getting bigger here,” he told reporters after the town hall meeting on Thursday morning. “We’ve got a game plan in place. We’ve stuck to it in spite of people saying it’s not working. We believe it’s working.” (30/12/2011)

Good campaign organization improves Santorum’s position in the primary elelctions


(48) Mr. Romney left Iowa on Friday afternoon for a quick overnight jaunt to New Hampshire. But as he campaigned with Gov. Chris Christie of New Jersey at a morning rally in Des Moines, Mr. Romney took a sharp jab at President Obama. (31/12/2011)

Romney’s arguments can harm Obama

A harms B

(49) Mr. Orban and Mr. Simor have sparred over bank decisions repeatedly, as Mr. Orban has pressed for policies that would spur faster economic growth. (16/12/2011)

Mr. Orban and Mr. Simor can harm each other with arguments

A harms B

(50) President Obama and his policy team have said that there is no greater threat to the fragile American recovery than a contracting and destabilized Europe. (29/11/2011)

Obama and his associates are united by the common goal of stabilizing the world economy

A and B are united by a common goal

(51) Mariano Rajoy was sworn in as prime minister on Wednesday, then hours later unveiled a slightly scaled-down government team that includes a former investment banker who will be tasked with cleaning up Spain’s finances and troubled banking sector. (22/12/2011)

Members of the Spanish government are united by their common political and economic goals

A and B are united by a common goal

Table 5. Emergent structures in SPORT metaphors: Additional examples

SPORT Metaphors 107

cognitive architecture that facilitates the processes of meaning construction. Furthermore, the form of the present emergent structure also stresses the role of spatial experience in human reasoning.43

Emergent structures in the form A harms B appeared in BOXING metaphors, and it is clear that this emergent structure has the same form as the most frequent emergent structure in CONFLICT metaphors. Such findings are perfectly plausible due to the fact that the BOXING frame is inherently linked to the frame of CONFLICT. However, it needs to be stressed that boxing involves a ‘controlled’ confrontation that takes place in accordance with a strict set of rules, suggesting that the participants’ ultimate goal is to simply win the game.

TEAM metaphors yielded a separate form of emergent structure: A and B are united by a common goal. This reflects the idea that team members are joined together by a common goal that is mutually beneficial for everyone, which is introduced into the blend via the borrowed compression from the TEAM frame. In the context of political discourse, it can be argued that by mirroring the collective cause of the political team onto its supporters, the supporters themselves are joined into a unique ‘virtual’ team, which can be used to bridge any differences and forge bonds between them (Silaški & Radić-Bojanić, 2010). In addition, such framing of political topics also presents an opportunity for provoking an emotional appeal with the supporters, since cheering for one’s team necessarily carries a certain degree of emotional involvement.

Additionally, it needs to be stressed that the above presented systematic forms of emergent structures can be linked to the highly entrenched nature of the conceptual key they represent. The same line of reasoning followed in the subsection dealing with the systematicity of emergent structures in CONFLICT metaphors applies here as well, along with all of the main conclusions presented in that section (see subsection 4.1.2.1.2 for details). Additional examples of the main forms of emergent structures in SPORT metaphors are given in Table 5.

4.1.2.2.3. Function and Emotional Appeal of SPORT Metaphors One of the main advantages of the use of the conceptual key POLITICS IS SPORT in the arena of politics

is that it creates the illusion of fair play, team spirit, and it introduces the idea that political players must follow a strict set of rules (Semino & Masci, 1996; Radić-Bojanić & Silaški, 2008; Silaški, Đurović, & Radić-Bojanić, 2009). Additionally, political “metaphors can stir emotions or bridge the gap between logical and emotional (rational and irrational) forms of persuasion” (Mio, 1997: 121), and they are also intended to simplify the political reality, thus rendering it more accessible to a wider audience (Mio, 1997). In this sense, by presenting the complex political issues in terms of sport, an illusion of simplicity is created, which marginalizes the notions of morality and values, while at the same time emphasizing the potential for provoking an emotional appeal (Charteris-Black, 2004; Ferrari, 2007). In other words, “within sport metaphors, the complexities of ideological and ethical issues are backgrounded and politics is presented as a relatively simple domain with clear participants [...], unproblematic goals [...] and unambiguous outcomes” (Semino & Masci, 1996: 250).

43 For example, such conclusions are aligned with Mandler, 2012.

SPORT Metaphors 108

Additionally, similarly to CONFLICT metaphors, SPORT metaphors also “equate physical power with political power, and they both share entailments portraying politics as a straightforward contest with an outright winner and clear ‘rules of engagement’” (Burnes, 2011: 2169).

For instance, talking about the presidential and primary elections as a race (examples 40-43), politicians and political groups as players (examples 44 and 45), political process as a game (examples 46 and 47), arguments between politicians as jabs or sparring (examples 48 and 49), and the political group as a team (examples 50 and 51), takes the focus of the discussion away from the actual essence and complexity of the political debate, and shifts it onto a far simpler concept of sport. This in turn conceals the persuasive aspect of the overarching entrenched conceptual key POLITICS IS SPORT, rendering supporters far more suitable for manipulation. Additionally, the “repetition of such metaphors results in dulling of the critical faculties rather than awakening them” (Edelman, 1964, cited in Mio, 1997: 119), which gives way for the installation of massive perception management mechanisms.

In short, it can be concluded that the presented examples of SPORT metaphors conform to the general mechanisms of persuasion proposed in Mio (1997), which include simplification, manipulation of underlying

symbolic representations, and the potential to provoke an emotional appeal. As argued in section 2.1.10, the fact that the present corpus consists of entrenched metaphors actually facilitates the persuasive aspect of metaphorical language in the political discourse of daily newspapers, since these entrenched structures operate mostly at an unconscious level.

4.1.2.2.4. Metaphor Systems and Metaphor Interaction SPORT metaphors also showed a certain degree of interaction, with 23.71% of metaphorical

expressions corresponding to the conceptual key of SPORT appearing in elaborate metaphor systems. As discussed in the case of CONFLICT metaphors, metaphor interaction in metaphor systems can serve to amplify the rhetorical potential of individual metaphors, where individual metaphors serve as resonators to each other. The same line of reasoning and the conclusions pertaining to the function and structure of metaphor systems that were presented in the section dealing with CONFLICT metaphors also apply here.

(52) The delayed start to the advertising war — which by this point in the last two presidential election cycles was already in full swing — has been one of the more unforeseen aspects of a campaign season that has had all the other trappings of a highly competitive race: wild poll fluctuations, a fight over coveted endorsements and spirited back-and-forth among the candidates. (The New York Times, 6/12/2011)

(53) Mr. Romney said he would “keep on battling” if he fell short in the first contests in Iowa, where he has stepped up his efforts in recent weeks, and New Hampshire, which had always seemed to be a secure launching pad for him but now seems to be in play. (The New York Times, 15/12/2011)

(54) Some quickly endorsed Gov. Rick Perry when he jumped into the race in August. (4/12/2011) (55) What his team did not expect was that they would be so cash-poor at this stage of the race. (4/12/2011) (56) The pointed comments suggested a new dynamic in the presidential primary race, with Mr. Paul as a new and

enticing target. (The New York Times, 24/12/2011)

SPORT Metaphors 109

(57) Mr. Obama’s aides acknowledge that a general election campaign against Mr. Gingrich would be “a different race,” one that would present a less predictable, more energetic and personally harder-hitting opponent than Mr. Romney would. (The New York Times, 8/12/2011)

Some additional examples of metaphor systems from the present corpus are listed above, and they

show interaction between different conceptual metaphors. Example 52 shows interaction between two CONFLICT metaphors and a SPORT metaphor; example 53 contains a metaphor system that includes two CONFLICT metaphors, one SPORT metaphor, and an additional CAUSED MOTION metaphor; example 54 shows an interaction between a structural SPORT metaphor and an ontological CONTAINMENT metaphor, where the race is viewed as a container; example 55 shows a system that contains two sport metaphors; examples 56 and 57 show interaction between a SPORT and a CONFLICT metaphor.

Network Optimization 110

4.1.2.3. Network Optimization in CONFLICT and SPORT Metaphors

In addition to the above discussed issues of vital relations, their compressions, the Governing

Principles for Compression, and the systematic nature of compression, another important aspect of meaning construction in single-scope metaphorical conceptual integration networks concerns network optimization. In that sense, and in line with the argument presented in Fauconnier and Turner (2002: 325-339), the present section will discuss the levels of optimization achieved in the previously described networks that were instances of CONFLICT and SPORT metaphors.

It is important to note that networks rarely achieve the maximum level of satisfaction when it comes to individual Optimality Principles. Instead, these principles usually compete, and the final network ‘works’ by being optimized for only some of these principles to various extents. For example, while single-scope networks are optimized for Topology, the Topology principle often clashes with some of the governing principles for compression. Namely, in the case of metaphorical single-scope networks, Topology is obviously satisfied between the blend and the source input that provides the organizing frame of the blend. Additionally, Topology is also satisfied between the blend and the target input, since the topological structures of the two input spaces are pre-aligned by the vital relation of Analogy that holds between their causal and intentional structures, and between their counterparts that are involved in the partial cross-space mappings, i.e. in the matching process. If a slight detour into the CMT domain is taken, such alignment between the topologies can be accounted for by the Invariance Principle. Namely, the fact that the target input image schematic structure is invariant to that of the source input suggests that the topologies of the two domains need to achieve a high degree of topological prealignment in order for a metaphoric link to be established between them. The CBT offers an additional step in this process, which is the formation of the blend. In that sense, the prealignment between the source and target inputs’ topologies is already accounted for by the former theoretical approach, while the latter framework offers an additional Topology principle that works over the entire network, securing the network’s coherence and cohesion.

Compression of one vital relation into another44 seems to be working directly against the Topology principle, since it creates novel vital relations in the blend by compressing the existing ones. However, this does not affect the frame-level topology of the blend, since the same casual and intentional structures that are analogous in the two inputs are also in force in the blend.45 In that sense, although “compression principles can resist Topology Principle’s preference for the preservation of topology” (Fauconnier & Turner, 2002: 328), the relevant topology, especially at the frame-level, remains preserved and the blend achieves an equilibrium between Compression and Topology. In other words, the final network is optimized for both compression and Topology to the extent that secures a ‘stable’, coherent blend at the human scale. 44 As discussed in the previous sections dealing with CONFLICT and SPORT metaphors, the most dominant compression of this type is that of Analogy into Uniqueness. 45 This is the general conclusion reached in the previous sections dealing with the analysis of CONFLICT and SPORT metaphors.


The Integration principle can also compete with Topology, since achieving an integrated blend carries an inherent necessity for compression. Still, in line with the previous discussion, the level of Integration needs to be aligned with the degree of optimization of the frame-level Topology. Without that, the network simply won’t work, i.e. the developed blend will be unstable. In that sense, a successful network needs to work out an acceptable level of satisfaction of these two principles in order for an equilibrium between them to be achieved. It is worth noting that the network optimization for these two principles is facilitated by the fact that these are instances of single-scope networks, without serious frame-level clashes between the inputs.

The Pattern Completion principle seems to work in concert with the Integration principle. Namely, the blend of a single-scope network inherits its organizing frame from the source input, and the blend’s topology is completed by additional background knowledge structures related to that organizing frame. In addition to strengthening the existing vital relations in the blend, these background knowledge structures also serve to reinforce the already integrated compressions. With that in mind, it can be argued that the Pattern Completion Principle and the Integration principle actually conspire to make the network function better. Namely, the process of Integration seems to be made both more coherent and cohesive by the process of Pattern Completion. This, however, may not be the case in a double-scope network, where the two principles might clash, owing to the fact that background knowledge structures related to the organizing frames of the inputs, which can be recruited in the blend, could in fact clash with some aspects of the topological organization of the novel organizing frame of the blended space in a double-scope network.

The main role of the Web principle is to make the network a coherent whole. In other words, the network needs to satisfy the Web principle in order for it to be optimized as a unit, which renders the process of meaning construction, i.e. compression and decompression of the blend, possible. In some cases, the Web principle “combines with Integration to force novel integrations in the blend” (Fauconnier & Turner, 2002: 332), which can facilitate some of the governing principles for compression, like creation by compression. Additionally, the fact that compressions in the blend have their counterparts in each of the inputs facilitates the optimization of the Web principle.

However, the Web principle can also compete with other principles, like Topology. Namely, the Topology principle restrains Web in the sense that “we must not project from the blend emergent topology that is inappropriate for the inputs” (Fauconnier & Turner, 2002: 331). As a result, optimization of the Web principle is directly governed by Topology. In the case of single-scope metaphorical networks this process is actually facilitated by the prealinged topological structures of the inputs, and the organizing frame of the blend that is adopted from the source input. In that sense, although Topology restrains the Web principle, the main idea behind the Web principle which suggests “that we should not disconnect valuable web connections to the inputs” (Fauconnier & Turner, 2002: 332) remains preserved.

Additionally, the Web principle could also be linked to the “molecule-model” discussed above (see Figure 15 for details). Namely, just as the Web principle accommodates the entire web of connections between


the blend and other mental spaces in the network, the on-line aspect of the “molecule-model” directly implicates the entire network of background compressions.

As discussed in Fauconnier and Turner (2002: 339), Unpacking in single-scope networks works similarly as in mirror networks. Namely, while the blend is well integrated at the frame-level, integration at lower levels is weaker. Additionally, the Unpacking process is conditioned by the immediate discourse context, i.e. “the Unpacking Principle is not purely one of structure within the network but, more broadly, one of communication, since the unpacking possibilities offered by the blended space will depend on what is already active in the context of communication” (Fauconnier & Turner, 2002: 333). In other words, successful Unpacking requires the construction of the fully developed discourse lattice.

Another important cue for Unpacking in the case of metaphorical single-scope networks are the links that exist between the compressed elements in the blend, and their counterparts in the input spaces. These connections that are directly influenced by Integration, Topology, and Web, present themselves as invaluable prompts for blend deconstruction, i.e. Unpacking. In light of these arguments, it can be concluded that Unpacking is directly constrained by Integration, Topology, and Web. This suggests that for the network to be optimized for Unpacking, it must first be optimized at least for the other three principles, which in turn reflects the notion that optimality principles need to be considered as a dynamic set of patterns that interact with each other as discourse unfolds. An additional prompt for Unpacking the blend can be found in compressions of one vital relation into another, i.e. in the compression of outer-space vital relation into inner-space vital relations, which provides a direct link between the compressions in the blend and their corresponding counterparts in the input spaces.

The Relevance principle is constrained by a set of other principles. Firstly, the interaction between Integration and Topology constrains the list of elements that will be relevant in the blend, most of all in the sense that the list will be the result of important outer-space compressions between the inputs. Secondly, Web also serves as an important constraint to this list. However, Web interacts with both Integration and Topology, while Integration interacts with Pattern Completion, and all of these interactions also constrain the possibilities of network optimization for Relevance. Thirdly, an additional factor in this interplay is the Pattern Completion principle which can serve to amplify the principle of the Intensification of Vital Relations in the blend via background knowledge recruitment, and therefore possibly render some elements more relevant than others. Finally, perhaps the most transparent limitation to Relevance is that imposed by the Unpacking principle, where the relevance of an element in the blend is determined by its role in Unpacking the blend. All these facts sanction a conclusion that the optimization of a given network for Relevance is actually constrained by its optimization for all other principles.

In summary, the previous discussion reflects the dynamic interplay of optimality principles that takes place in the process of meaning construction in single-scope metaphorical networks. Regardless of whether some of the principles stand in direct opposition, or support each other, it is important to foreground the nature


of their interaction and the fact that the optimization of the final network requires that an equilibrium be established between these principles, thereby satisfying each of them to a certain extent. Furthermore, the role of the immediate discourse context must not be neglected, since the fully developed discourse lattice can also serve as an important constraint.


4.2. Present Research: Part 2

The methodology and theoretical framework used in the second part of the present research were adopted from research on affect and emotions. This section of the paper involved empirical research via appropriate questionnaires based on the theoretical framework introduced in section 2.3. The specific structure of these questionnaires will be described in the following paragraphs.

The study included a total of 101 participants, 76 females, and 25 males, with the participants’ average age of 22.68 years-old (Std. Deviation of 1.456). The age range was 21 – 27. All participants were students of English at the Department of English, Faculty of Philosophy, University of Niš, Serbia, and the study included 44 third-year students, 31 fourth-year students, and 26 graduate students. The questionnaires were distributed to the participants during their lectures at the Faculty of Philosophy, University of Niš, and they were given thirty minutes to complete them.

Based on the theoretical framework discussed in section 2.3, the study included two types of questionnaires. Type-one questionnaires (Questionnaires 1 and 3)46 were based on the Russell’s circumplex model of affect (Russell, 1980; 1989), and they were designed to measure valence, and arousal, the two components of affect. Questionnaire 1 was designed to measure the response to metaphorical expressions corresponding to the conceptual key POLITICS IS CONFLICT, while Questionnaire 3 was designed to measure the reaction to metaphorical expressions corresponding to the conceptual key POLITICS IS SPORT. In line with the methodology used in Nešić et al. (2009) and Nešić et al. (2010), the participants were presented with a 10-point Likert scale, ranging from 0 to 9, on which they were instructed to record their affective responses to the target stimuli.47 Type-one questionnaire contained eight sentences extracted from the corpus, and in each sentence there was an expression (i.e. a metaphorical expression) that was bolded and underlined. Additionally, before being included in the questionnaire, each sentence containing a metaphorical expression (i.e. a target stimulus) was filtered for potentially biased information. Namely, all of the personal names, names of political fractions, and similar information were excluded and replaced with appropriate pronouns, or other neutral phrases. The experimental subjects were instructed to read each sentence carefully, and then decide on how pleasant (in the sense of valence), and how exciting (in the sense of arousal) the emphasized expression was for them, and rate their answers on the previously described Likert scale. The subjects were not primed, i.e. they were not told that the emphasized expressions represented metaphorical expressions corresponding to CONFLICT and SPORT metaphors.

The main aim of type-one questionnaires was to determine whether there was an actual affective response in terms of valence and arousal to the target stimuli, i.e. the selected metaphorical expressions included in the questionnaire. Judging by the scale construction, with “0” corresponding to the most unpleasant 46 See Appendix A and Appendix C for the full content of Questionnaires 1 and 3, respectively. 47 In terms of valence: 0 = the least pleasant, 9 = the most pleasant; in terms of arousal: 0 = the least arousing, 9 = the most arousing.


reaction (in terms of valence), or to the least arousing reaction (in terms of arousal), and “9” to the most pleasant reaction (in terms of valence), or the most aroused reaction (in terms of arousal), it is obvious that the arithmetic middle of the scale that corresponds to the value of “4.5”, would describe an ideally neutral reaction. In order to make sure that the reactions of experimental subjects were different from neutral, mean values of valences and arousals for each individual stimulus respectively were compared against the calculated mean neutral value of 4.5. The aim of such a procedure was to establish whether the mean values of valences and arousals showed statistically significant differences (p<.05) from the value that would describe an ideally neutral stimulus. This was achieved through the use of One-Sample T-Tests.48

Type-two questionnaires (Questionnaires 2 and 4)49 were based on an emotion concepts checklist, similar to the ones described in Watson, Clark, and Tellegen (1988), and Thompson (2007). The present checklist contained four positive adjectives (happy, proud, motivated, and supportive), and four negative adjectives (sad, angry, disappointed, and alarmed), which stood for the emotion concepts commonly used to describe political discourse. Such a closed adjective checklist was meant to create a controlled response

environment. In other words, the participants had to decide between the adjectives that were offered, and were not allowed to provide additional descriptions of their own. The main motivation for such an approach was the desire to obtain systematic results on a clearly defined closed scale. Questionnaire 2 was designed to measure the response to metaphorical expressions corresponding to the conceptual key POLITICS IS CONFLICT, while Questionnaire 4 was designed to measure the reaction to metaphorical expressions corresponding to the conceptual key POLITICS IS SPORT.

Type-two questionnaires contained the same stimuli that were used in their corresponding type-one questionnaires. Namely, since Questionnaires 1 and 3 were both designed to measure the subjects’ reactions to metaphorical expressions corresponding to the conceptual key POLITICS IS CONFLICT, these two questionnaires contained the same stimuli. On the other hand, Questionnaires 2 and 4 were designed to measure the subjects’ responses to metaphorical expressions corresponding to the conceptual key POLITICS IS

SPORT, and these two questionnaires also contained the same stimuli. After each sentence with a metaphorical expression, participants were presented with a list of eight

previously mentioned emotion concepts, i.e. adjectives, and were instructed to describe their emotional reaction to a given stimulus in terms of the presented adjective checklist. Each adjective was followed by a 10-point Likert scale (0-9), on which the participants rated the degree to which an adjective described their emotional response to the given stimulus.50 All questionnaires were distributed to the participants on the same occasion.

48 The statistical analyses were performed using the SPSS software, version 20 (Statistical Package for Social Sciences). 49 See Appendix B and Appendix D for the full content of Questionnaires 2 and 4, respectively. 50 0 = the given adjective does not describe the emotional reaction at all; 9 = the given adjective describes the emotional reaction completely.

Questionnaires 1 and 2 116

Based on the results obtained from the type-two questionnaires, mean values for each of the adjectives describing each metaphorical expression respectively were calculated. Comparison of mean values was used to determine which of the adjectives had the maximum mean value, i.e. which adjective proved to be the best choice for describing the given target stimulus. In the next step, the maximum mean value was compared to the means of the remaining seven adjectives to determine whether there was a statistically significant difference between them. If the measured p value was lower than .05, then there was a statistically significant difference, and the adjective with the highest mean value was a better choice for describing the reaction to a given stimulus. If p was greater than .05, then there was no statistically significant difference between the two mean values, and both adjectives were dubbed equally good for describing the reaction to the given stimulus. This part of the analysis was performed using the One-Sample T-Test.51

4.2.1. Questionnaire 1: Results and Discussion

Questionnaire 1 was a type-one questionnaire, and it contained metaphorical expressions, i.e. the target stimuli corresponding to the conceptual key POLITICS IS CONFLICT. 4.2.1.1. Scale Reliability

The reliability of the scale used in the first questionnaire was determined based on Cronbach’s Alpha, and it was calculated to be .882. Twelve participants (i.e. 11.9%) were excluded from the calculation on account that the outliers were eliminated from the database in the previous step. Cronbach’s Alpha value calculated prior to the deletion of the outliers was calculated to be .894, which suggests that owing to the elimination of the outliers, Cronbach’s Alpha value was reduced by 1.34%, which does not significantly affect the scale reliability.

Inspection of the Inter-Item Correlation Matrix showed consistent positive values of inter-item correlation throughout the matrix, which suggests that “all items are measuring the same underlying characteristic” (Pallant, 2011: 100). Additionally, values of the Corrected Item-Total Correlation were consistently higher than the marginal value of .3 (ibid.: 100), and ranged from .383 to .672. Values of Cronbach’s Alpha if Item Deleted ranged from .871 to .882, which suggests that excluding any of the variables would either not affect, or would reduce the reliability of the scale.

51 The statistical analyses were performed using the SPSS software, version 20 (Statistical Package for Social Sciences).


4.2.1.2. Assumptions Required for the Application of Parametric Tests 52 In line with the methodology proposed in Pallant (2011: 205-207), the following assumptions were

tested: 1) Level of Measurement. All variables were measured on a continuous scale, suggesting that this

assumption was satisfied. 2) Random Sampling. Although the sample used in the research was not a proper random sample, since

the research explicitly targeted students of English, it can still be considered a good approximation of random sampling since the actual experimental subjects were picked randomly within a closed group of potential participants. Therefore, it can be concluded that the random sampling assumption was still partially satisfied.

3) Independence of Observations. This assumption was satisfied, since the measurements were not co-dependent.

4) Normal Distribution. In line with the general methodological procedures presented in Pallant (2011, 59-65), the data normality was assessed based on the following factors: 5% trimmed mean, skewness,

kurtosis, the Kolmogorov-Smirnov test, and the visual inspection of the histograms and boxplots. All of the cases were valid, i.e. none of the cases was excluded from this analysis.53

The comparison of mean values to their corresponding 5% trimmed mean values showed notable differences in a number of cases (metaphor 1 valence, metaphor 2 valence, metaphor 3 valence, metaphor 6 valence, and metaphor 7 arousal). Additionally, the visual inspection of the histograms, and the results presented in the boxplots also suggested that there were certain outliers. Consequently, these extreme values were eliminated from the database, and their slots remained empty. The visual inspection of the histograms also suggested that the data distribution was similar to a normal distribution, with certain offsets to the left and right. Still, these offsets were within the acceptable limits.

The measurements of skewness showed values within the [-1,+1] range, which are suggested to be acceptable for a normal distribution (Dancey & Reidy, 2007; Cronk, 2006; Pallant, 2011). Additionally, four samples showed negative skews, i.e. their distribution was shifted to the right, while the remainder of the samples showed positive skews. Consistent negative values of kurtosis suggest that all samples showed platykurtic normal distribution (Dancey & Reidy, 2007: 79). The Kolomogrov-Smirnov test was also consistent, and showed p values lower than .05. Although this might suggest that the assumption of a normal distribution was violated, such results are to be expected in studies with a large number of experimental subjects (Pallant, 2011: 63). In conclusion, the assumption of a normal distribution was satisfied.

52 A detailed overview of results pertaining to these requirements will be presented only for Questionnaire 1. The remaining three questionnaires have also been analyzed in relation to these requirements, and based on the obtained results it was concluded that all questionnaires satisfy the assumptions required for the application of parametric tests. 53 For potential cases of missing values the “exclude cases pairwise” option was chosen.


Based on the above presented data, it can be concluded that all of the assumptions required for the parametric tests to be performed were satisfied.

4.2.1.3. Means and One-Sample T-Test

Based on the results obtained from the statistical analysis, the following conclusions were reached: 1) The mean values of valences for all tested metaphorical expressions show a statistically significant

difference from the calculated critical mean value of 4.5 that would correspond to a neutral stimulus. Additionally, the mean values of all valences are gravitating towards lower values, suggesting that the stimuli were assessed as mostly unpleasant, with the exception of metaphors 5 and 8 which were categorized as somewhere in between moderately unpleasant, and neither pleasant nor unpleasant.

2) The mean values of arousals showed a tendency ranging from neutral to not so exciting. Namely, arousals for metaphors 3, 4, 6, and 8 did not show any statistically significant differences compared to the neutral stimulus. Therefore, the four previously mentioned stimuli were rated as neither arousing, nor non-arousing. The mean values of metaphors 1, 2, 5, and 7, on the other hand, showed a statistically significant difference compared to a neutral stimulus. The mean values of arousals of these metaphors also showed a very consistent distribution. Namely, all of their values are situated within a narrow range between 3.65 and 3.92, which suggests that the stimuli were rated as moderately non-arousing.

Valence Arousal One-Sample T-Test One-Sample T-Test

Stimulus Mean

Test Value = 4.5 Mean

Test Value = 4.5 1) nominating battle 3.32 .000 3.74 .002

2) re-election fight 3.03 .000 3.88 .016

3) barrage of attacks 2.41 .000 4.07 .091

4) the showdown 3.88 .009 4.65 .525

5) to defend 4.04 .043 3.92 .008

6) who could defeat 3.04 .000 4.08 .070

7) taking aim at 3.22 .000 3.80 .002

8) (war) front 3.99 .026 4.08 .065

Table 6. Questionnaire 1 - Results

In short, the one-sample t-test showed that the majority of variables (all valence variables, and four out of eight arousal variables) showed statistically significant differences from the values expected for an ideally neutral stimulus. Therefore, it can be concluded that the target stimuli, i.e. the selected metaphorical expressions included in the questionnaire, provoked an affective response with the experimental subjects predominantly in terms of valence. The only two potentially disputable cases are metaphors 5 and 8, where valences 5 and 8 have p values of .043, and .026, respectively, which are a lot higher than the obtained p values for other valences.


Still, the present findings obtained from the one-sample t-test sanction the previously stated claim that the majority of target stimuli did provoke an actual affective response with the participants in the study, which in turn lends reasonable support to the theoretically derived conclusions from the first part of the present research that dealt with the theoretical corpus analysis. Additionally, it is also worth noting that the dominant role in the construction of the affective experience was played by the valence component of affect. Such findings are in line with the arguments presented in Barrett (2006) and Russell (2003), where it was suggested that in using self-report measurements of affect, the dominant response is recorded on the scale between pleasant-unpleasant (i.e. in terms of valence), whereas the arousal component is more difficult to record through the use of self-reports. 4.2.2. Questionnaire 2: Results and Discussion

Questionnaire 2 was a type-two questionnaire, and it contained the same metaphorical expressions, i.e. the same target stimuli as Questionnaire 1. 4.2.2.1. Scale Reliability

The scale reliability was assessed primarily based on the value of Cronbach’s Alpha. Inter-Item Correlation Matrix, values of Cronbach’s Alpha if Item Deleted, and Corrected Item-Total Correlation values were also analyzed to substantiate the initial findings obtained from Cronbach’s Alpha.

The calculated value of Cronbach’s Alpha was .972. However, due to the elimination of the outliers performed in the previous step, 26 participants, i.e. 25.7% of participants were excluded from the calculation. The value of Cronbach’s Alpha prior to the elimination of the extreme values was .968, suggesting that the scale reliability improved by 0.41% after the outliers were removed. This information suggests that Cronbach’s Alpha value remained very high, well above .95 in both cases, so that the scale reliability was not significantly affected due to the elimination of the outliers.

Inspection of the Inter-Item Correlation Matrix revealed only 35 instances of negative values, i.e. only 0.85% of cases, suggesting a very high degree of inter-item correlation. Values of Cronbach’s Alpha if Item Deleted ranged from .971 to .972. This means that the deletion of an item would either leave Cronbach’s Alpha unchanged, or it would reduce it by 0.103%. Additionally, the Corrected Item-Total Correlation values ranged from .390 to .735, with 62 out of 64 variables, i.e. 96.875% of variables achieved values either above, or well above .4. This was higher than the marginal value of .3 suggested in Pallant (2011: 100). The only exception was the variable ‘metaphor 5 disappointed’, which achieved the value of .291.

In sum, based on the discussion in the previous paragraphs, the scale used in the second questionnaire showed a high degree of reliability and consistency.


4.2.2.2. Means and One-Sample T-Test Based on the statistical analysis, the following conclusions were reached (see Tables 7 and 8):

1) the emotional reaction to the first stimulus can be described equally well by the adjectives motivated,

supportive, and alarmed; 2) the reaction to the second stimulus can be described by the adjectives motivated and supportive; 3) the reaction to the third stimulus is best described by the adjectives angry and alarmed;

Stimulus Means; Standard Deviations; One-Sample T-Tests

Adjectives happy motivated proud supportive sad angry disappointed alarmed Means 1.53 3.03 2.33 3.00 1.28 2.05 1.51 2.52 SD 1.787 2.575 2.600 2.665 1.865 2.632 2.145 2.678

1) nominating battle

One-sample T-test (test value: 3.03)

.000 / .008 .910 .000 .000 .000 .061


2) re-election fight


.000 / .002 .352 .000 .000 .000 .010


3) barrage of attacks


.000 .001 .000 .000 .000 .144

.000 /

Adjectives happy motivated proud supportive sad angry disappointed alarmed Means 2.22 3.07 2.33 2.79 .89 1.55 1.13 2.37 SD 2.444 2.909 2.627 2.776 1.215 2.139 1.664 2.533

4) the showdown


.001 / .005 .317 .000 .000 .000 .006

Adjectives happy motivated proud support. sad angry disappointed alarmed Means 2.30 2.95 2.69 3.19 1.43 1.83 1.83 2.89 SD 2.663 2.951 2.863 3.058 1.861 2.213 2.413 2.922

5) to defend


.001 .417 .084 / .000 .000 .000 .306


6) who could defeat


.003 / .063 .347 .000 .000 .001 .220


7) taking aim at


.000 .009 .000 .000 .000 .035 .000 /

Adjectives happy motivated proud support. sad angry disappointed alarmed Means 1.59 2.41 2.14 3.06 1.23 2.11 2.04 2.49 SD 1.968 2.357 2.490 2.615 1.686 2.679 2.596 2.678

8) (war) front


.000 .007 .000 / .000 .001 .000 .033

Table 7. Questionnaire 2: Means, Standard Deviations, One-Sample T-Tests


4) the emotional response to the fourth stimulus can be best described by the adjectives motivated and supportive, as was the case with the second stimulus. This similarity can be further accounted for by the relatedness between the FIGHT and SHOWDOWN frames;

5) the reaction to the fifth stimulus can be described by the adjectives motivated, proud, supportive, and alarmed;

6) the response to the sixth stimulus can de described by the same group of adjectives as was the case with the fifth stimulus, i.e. motivated, proud, supportive, and alarmed;

7) the response to the seventh stimulus is best described by the adjective alarmed; 8) the reaction to the eighth stimulus can be described by the adjective supportive.

It can be concluded that all experimental subjects were consistent in qualitatively describing their emotional responses to the presented stimuli, since their descriptions showed a consistent dominance of three adjectives from the list, motivated, supportive, and alarmed, while the adjective proud appeared only two times, and the adjective angry appeared only on one occasion (see Tables 7 and 8 for details). Additionally, it should also be noted that only the three most dominant adjectives, i.e. motivated, supportive, and alarmed appeared with the maximum mean values, while the adjectives angry and proud never presented themselves as the best choice for describing any of the stimuli (see Table 7 for details).

Stimulus Adjectives 1) nominating battle motivated, supportive, alarmed

2) re-election fight motivated, supportive

3) barrage of attacks angry, alarmed

4) the showdown motivated, supportive

5) to defend motivated, proud, supportive, alarmed

6) who could defeat motivated, proud, supportive, alarmed

7) taking aim at alarmed

8) (war) front supportive

Table 8. Questionnaire 2 – Final Results Such results suggest that all metaphorical expressions representing instances of CONFLICT metaphors

provoked a relatively uniform emotional response in terms of quality with all of the participants in the study. This in turn supports the initial theoretically derived assumption that conceptual metaphors in political discourse can in fact provoke an actual emotional reaction with the audience, i.e. readers. However, bearing in mind that the participants were presented with a 10-point Likert scale (0-9), while the highest recorded mean value for any of the adjectives was 3.19, it can also be concluded that none of the adjectives were too suitable for describing the actual quality of the emotional response. Still, a closed adjective checklist represents a controlled testing environment, and the fact that the obtained results show high degree of consistency with all


experimental subjects licenses a valid conclusion that there in fact was an actual emotional response to the target stimuli.

Additionally, for purposes of further research, a revised adjective checklist could yield even more convincing results. Methodologically-wise, the scale used in the second questionnaire could benefit from a factor-analytic approach, where an initial list containing a larger number of adjectives would be reduced to a lower number, which can be achieved following the methodological procedures described in Pallant (2007: 185-201). Consequently, this would also require a greater number of experimental subjects. 4.2.3. Questionnaires 1 and 2: Comparison of Findings

In general terms, the mean values of valences corresponding to mostly unpleasant reactions, coupled with the mean values of arousals corresponding to neutral-to-mostly-non-arousing activation and not so arousing activation, provoked the feelings of motivation, pride, support, anger, and alarm (see Table 9 for details).

Categorization of affective reactions through the feeling of alarm and anger can be directly linked to the negative valences and neutral-to-negative arousals. Categorization of affective reactions in terms of motivation, pride, and support can be explained as being caused by the desire to overcome the tension introduced by the various frames of conflict presented by the stimuli. In other words, the fact that the participants categorized their affective reactions with negative valences and neutral-to-negative arousals through the feelings of motivation, pride, and support suggests that their emotional reaction was aligned with the intentional and causal structures of the blended spaces from the corresponding metaphorical integration networks. For instance, in stimulus 2, the participants experienced motivation and support which arise from the fight frame; the same reaction was recorded in stimulus 4, where these emotional responses arise from the showdown frame, etc.

In short, the emotional experience described in qualitative terms in the second questionnaire can be linked to the affective reaction described in terms of valence and arousal in the first questionnaire. Additionally, cognitive mechanisms such as the recruitment of background knowledge seem to both aid, and facilitate the process of affect categorization, which is in line with the arguments presented in Barrett (2006), Lindquist and Barrett (2008), and Wilson-Mendenhall et al. (2011). In that sense, the conclusions pertaining to the potential of conceptual metaphors to provoke an emotional response in the context of political discourse, reached based on the analyses in the CBT framework, seem to be substantiated by the findings from the first two questionnaires. Furthermore, apart from mere theoretical predictions, CBT also appears to be a viable tool that can serve to shed more light on the process of affect categorization into specific emotions, owing to the points of intersection between the Conceptual Act Theory of Emotion and the CBT model, discussed in section 2.3.4.


Stimulus Questionnaire 1 Questionnaire 2 1) nominating

battle mean valence: 3.32 (p=.000) mean arousal: 3.74 (p=.002)

moderately unpleasant and non-arousing

motivated, supportive, alarmed

2) re-election fight

mean valence: 3.03 (p=.000) mean arousal: 3.88 (p=.016)

unpleasant and moderately non-arousing

motivated, supportive

3) barrage of attacks


very unpleasant and neither arousing nor non-arousing

angry, alarmed

4) the showdown

mean valence: 3.88 (p= .009) mean arousal: 4.65 (p=.525)

moderately unpleasant and neither arousing nor non-arousing


5) to defend mean valence: 4.04 (p=.043) mean arousal: 3.92 (p=.008)

moderately unpleasant to neither unpleasant nor pleasant, and moderately non-arousing

motivated, proud, supportive, alarmed

6) who could defeat


unpleasant and neither arousing nor non-arousing

motivated, proud, supportive, alarmed

7) taking aim at


unpleasant and moderately non-arousing

alarmed

8) (war) front mean valence: 3.99 (p=.026) mean arousal: 4.08 (p=.065)

moderately unpleasant to neither unpleasant nor pleasant, and neither arousing nor non-arousing

supportive

Table 9. Comparison of Results from Questionnaires 1 and 2 Namely, processes of schema induction (in the sense of Evans & Green, 2006) can account for the

influence of background knowledge on the construction of emotional experience, whereas the earlier proposed backward projections and the coupled elaboration of the blend can be used to reflect the influences of the immediate discourse context on the categorization of the initial affective reaction into a specific emotional response. Again, these ideas are aligned with the notion of emotions being a product of categorization (e.g. Russell, 1980; 2003), and the Conceptual Act Theory of Emotion (Barrett, 2006). Additionally, these conclusions are also in accord with the notion of situated conceptualizations (Barsalou, 2009; Wilson-Mendenhall et al., 2011), where entrenched situated conceptualizations correspond to the notion of highly specified entrenched frames available for recruitment when prompted by appropriate contextual cues.


Consequently, just as when it comes to meaning construction, emotions are also not to be understood as closed, strictly defined sets, but rather as products of the process of emotion construction which involves the categorization of more abstract affective states and the activation of the conceptual knowledge about emotions, which is also influenced by the background knowledge structures. Additionally, the fact that most of the stimuli, i.e. metaphorical expressions, provoked reactions described in terms of multiple adjectives further reinforces the claim that emotions are fuzzy sets, and that the boundaries between them are not clear-cut (e.g. Russell, 1980; Niedenthal, 2008).

4.2.4. Questionnaire 3: Results and Discussion

Questionnaire 3 was a type-one questionnaire, and it contained metaphorical expressions, i.e. the target stimuli corresponding to the conceptual metaphor POLITICS IS SPORT. 4.2.4.1. Scale Reliability

Cronbach’s Alpha value for the third questionnaire was calculated to be .924. Four participants were excluded from the calculation because the outliers were eliminated from the database in the previous step. Cronbach’s Alpha value calculated before the deletion of the outliers was .925, suggesting that the deletion of the outliers reduced the scale reliability by mere .1081%.

Elements of the Inter-Item Correlation Matrix showed consistent positive values throughout the entire matrix, which suggests that “all items are measuring the same underlying characteristic” (Pallant, 2011: 100). In addition, values of the Corrected Item-Total Correlation were consistently higher than the marginal value of .3, and ranged from .466 to .800. Values of Cronbach’s Alpha if Item Deleted ranged from .915 to .924, suggesting that the deletion of an item would either reduce, or not affect the reliability of the present scale. 4.2.4.2. Means and One-Sample T-Test

This part of the analysis employed the general methodology for type-one questionnaires. Based on the obtained results, the following can be concluded: 1) The mean values of valences for stimuli 1, 2, 3, 5, 6, 7, and 8 showed statistically significant differences

from the critical value of 4.5, whereas stimulus 4 did not show such difference. Additionally, the mean valences of stimuli 1, 2, 3, 5, 6, and 7 were situated in the negative, i.e. the unpleasant part of the scale. The only stimulus with the mean valence in the positive, i.e. pleasant part of the scale was stimulus 8 (team).

2) The mean values of arousals for stimuli 3, 5, and 6 showed statistically significant differences from the critical value of 4.5, and these means were placed in the unpleasant part of the scale. Stimuli 1, 2, 4, 7,


and 8, on the other hand, did not show such differences, and the recorded reaction to these stimuli can be dubbed neutral in terms of arousal.

Valence Arousal One-Sample T-Test One-Sample T-Test

Stimulus Mean

Test Value = 4.5 Mean

Test Value = 4.5 1) break out of the pack 3.09 .000 4.67 .498

2) last-minute sprints 3.37 .000 4.56 .828

3) presidential race 2.94 .000 3.67 .005

4) took the lead 4.54 .869 4.77 .342

5) political players 3.03 .000 3.21 .000

6) late in the game 3.52 .000 3.77 .005

7) sparring 3.09 .000 4.14 .190

8) team 5.36 .002 4.56 .810

Table 10. Questionnaire 3 - Results

In sum, it can be concluded that the dominant affective response was recorded in terms of valence, as was the case with CONFLICT metaphors. Reactions to stimuli 3, 5, and 6 were described as unpleasant, and non-arousing; reactions to stimuli 1, 2, and 7 were described as unpleasant in terms of valence, and neutral in terms of arousal; reaction to stimulus 8 was rated as pleasant in terms of valence, and neutral in terms of arousal; stimulus 4 (took the lead) did not show any activation in terms of valence and arousal, and therefore, the affective response to this stimulus was described as completely neutral. Based on such results, it can be concluded that, with the exception of stimulus 4 (took the lead), all other metaphorical expressions provoked a certain degree of an affective response with the participants, which was achieved predominantly in terms of valence. 4.2.5. Questionnaire 4: Results and Discussion

Questionnaire 4 was a type-two questionnaire, and it contained the same metaphorical expressions, i.e. the same target stimuli as Questionnaire 3. 4.2.5.1. Scale Reliability

Scale reliability was primarily assessed based on the values of Cronbach’s Alpha. The Inter-Item Correlation Matrix, values of Cronbach’s Alpha if Item Deleted, and the Corrected Item-Total Correlation values were also analyzed to substantiate the initial findings obtained from Cronbach’s Alpha.

Cronbach’s Alpha value was calculated to be .974. However, owing to the elimination of the outliers in the previous step, 51 participants were excluded from the calculation based on the listwise deletion. The value of Cronbach’s Alpha prior to the deletion of the outliers was calculated to be .967, which suggests that the


scale reliability was increased by .719% (<1%). This suggests that Cronbach’s Alpha remained very high in both cases, i.e. the deletion of the outliers did not significantly affect the reliability of the scale.

Inter-Item Correlation Matrix showed only 26 negative values, i.e. only .635% of cases (<1%), which suggests a high degree of inter-item correlation. Corrected Item-Total Correlation values ranged from .324 to .820, which was consistently above the marginal value of .3 suggested in Pallant (2007: 100). Values of Cronbach’s Alpha if Item Deleted ranged from .973 to .974, which suggests that the deletion of any of the items would either not affect the scale reliability, or would reduce it by mere .103%. 4.2.5.2. Means and One-Sample T-Test

This section of the analysis followed the general methodological guidelines used in the analysis of type-two questionnaires. Based on the calculated mean values, and the performed statistical tests (One-Sample T-Tests), the following conclusions were reached (see Tables 11 and 12): 1) the emotional reaction to the first stimulus can be described equally well by the adjectives happy,

motivated, proud, supportive, and alarmed; 2) the response to the second stimulus can be described by the adjectives motivated and supportive; 3) the response to the third stimulus can be described by the adjective alarmed; 4) the emotional reaction to the fourth stimulus is best described by the adjectives happy, motivated, proud,

and supportive; 5) the reaction to the fifth stimulus can be described by the adjectives angry, disappointed, and alarmed; 6) the response to the sixth stimulus can be described by the following set of adjectives: happy, motivated,

supportive, and alarmed; 7) the reaction to the seventh stimulus is best described by the adjectives supportive and alarmed; 8) the emotional response to the eighth stimulus can be described by the adjectives happy, motivated, proud,

and supportive. Questionnaire 4 showed a certain degree of systematicity in results, similar to the case of

Questionnaire 2. Namely, the four dominant adjectives used to describe the emotional response to the given stimuli were happy, motivated, supportive, and alarmed. In addition to these four adjectives, the adjective proud appeared three times, while the adjectives angry and disappointed appeared only once (see Tables 11 and 12 for details). Still, the present findings suggest that the selected metaphorical expressions corresponding to the conceptual key POLITICS IS SPORT did show an actual emotional activation recorded via the above-described adjective checklist. Additionally, as already noted in the section dealing with Questionnaire 2, the present adjective checklist serves as a controlled response environment, and the present questionnaire could also benefit from a revised adjective checklist developed through a factor-analytic approach.


Stimulus Means; Standard Deviation; One-Sample T-Test


1) break out of the pack


.111 / .116 .504 .000 .000 .000 .275

Adjectives happy motivated proud supportive sad angry disappointed alarmed Means 2.26 2.96 2.36 2.79 .54 .82 .66 1.69 SD 2.637 2.898 2.866 2.974 1.007 1.336 1.024 20221

2) last-minute sprints


.009 / .037 .572 .000 .000 .000 .000


3) presidential race


.000 .003 .000 .002 .000 .000 .000 /

Adjectives happy motivated proud supportive sad angry disappointed alarmed Means 3.36 3.59 3.14 3.26 .45 .88 .65 2.03 SD 2.955 3.163 3.052 3.004 .841 1.431 1.036 2.497

4) took the lead


.429 / .144 .275 .000 .000 .000 .000

Adjectives happy motivated proud supportive sad angry disappointed alarmed Means 1.35 1.40 .89 1.50 .92 2.05 2.05 2.18 SD 2.086 1.943 1.463 2.067 1.397 2.651 2.670 2.700

5) political players


.000 .000 .000 .001 .000 .622 .624 /


6) late in the game


.362 / .000 .250 .000 .004 .021 /


7) sparring


.000 .009 .001 .058 .000 .000 .000 /

Adjectives happy motivated proud support. sad angry disappointed alarmed Means 4.00 4.16 3.83 4.36 .52 .36 .27 .76 SD 2.929 2.962 3.033 3.012 .962 .735 .634 1.259

8) team


.220 .496 .083 / .000 .000 .000 .000

Table 11. Questionnaire 4: Means, Standard Deviations, and One-Sample T-Tests


Stimulus Adjectives 1) break out of the pack happy, motivated, proud, supportive, alarmed

2) last-minute sprints motivated, supportive

3) presidential race alarmed

4) took the lead happy, motivated, proud, supportive

5) political players angry, disappointed, alarmed

6) late in the game happy, motivated, supportive, alarmed

7) sparring supportive, alarmed

8) team happy, motivated, proud, supportive

Table 12. Questionnaire 4 - Results 4.2.6. Questionnaires 3 and 4: Comparison of Findings

Table 13 provides a parallel overview of the results obtained from Questionnaires 3 and 4. The emotional reactions described in terms of adjectives alarmed, angry, and disappointed that appear in stimuli 1, 3, 5, 6, and 7, can be directly linked to the negative valences, i.e. the unpleasant activation in terms of valence. The emotional reactions described in terms of adjectives happy, motivated, proud and supportive recorded with stimuli 1, 2, 4, 6, 7, and 8 cannot be directly linked to negative valences, since the aforementioned adjectives are marked by a positive valence. Still the recorded responses are perfectly plausible, and they can be linked to background knowledge structures introduced by the immediate context, and the stimuli themselves (Barrett, 2006; Barsalou, 2009). In that sense, it can be argued that the construction of the emotional experience does not include the categorization of the more abstract affective states alone, but also involves the recruitment of background knowledge which influences the entire process, and such line of reasoning is in accordance with the arguments presented in Barrett (2006), Lindquist and Barrett (2008), Barsalou (2009), and Wilson-Mendenhall et al., (2011). Additionally, such conclusions are especially supported by the notion of entrenched situated conceptualizations available for recruitment in the process of emotion construction (Barsalou, 2009).

If we briefly revisit the ideas presented in Fauconnier and Turner (2008), it is obvious that the notions of cobbling and sculpting that the authors apply to the metaphorical conceptual integration networks are equally applicable to other aspects of human experience. Namely, it is obvious that some aspects of the process of emotion construction are developed explicitly on-line, while certain aspects of this process involve the recruitment of background knowledge structures in the form of highly specified entrenched frames via the process of schema induction (in the sense of Evans and Green, 2006). Consequently, both of these aspects facilitate and fuel the activities that are taking place on-line.


Stimulus Questionnaire 3 Questionnaire 4 1) break out

of the pack mean valence: 3.09 (p=.000) mean arousal:4.67 (p=.498)


happy, motivated, proud, supportive, alarmed

2) last-minute sprints

mean valence:3.37 (p=.000) mean arousal: 4.56 (p=.828)



3) presidential race


very unpleasant and non-arousing

alarmed

4) took the lead


neutral both in terms of valence and arousal

happy, motivated, proud, supportive

5) political players

mean valence:3.03 (p=.000) mean arousal: 3.21 (p=.000)

unpleasant and non-arousing

angry, disappointed, alarmed

6) late in the game


unpleasant and non-arousing

happy, motivated, supportive, alarmed

7) sparring mean valence: 3.09 (p=.000) mean arousal: 4.14 (p=.190)


supportive, alarmed

8) team mean valence: 5.36 (p=.002) mean arousal: 4.56 (p=.810)

PLEASANT and neutral in terms of arousal

happy, motivated, proud, supportive

Table 13. Comparison of Results from Questionnaires 3 and 4 The only stimulus that showed positive valence was stimulus 8 (team), where it can be argued that the

emotional reaction described by the adjectives happy, motivated, proud, and supportive is both a direct result of the categorization of positive valence, and the recruitment of the background knowledge structures related to the notion of a team. This suggests that the ideas of team spirit and joint effort to achieve the common goal are both pleasant in terms of affect, and described by positive emotional responses. In addition, it is worth noting that so far, this was the only stimulus rated with positive valence. All the other stimuli from all four questionnaires were rated by valences that were either unpleasant or close to neutral.

Stimulus 7 (sparring) was described in terms of the negative, i.e. unpleasant valence, neutral arousal, and the adjectives supportive and alarmed, which corresponds to the results obtained from Questionnaires 1 and 2. Namely, although sparring is essentially a boxing, i.e. a sport concept, it can also be related to the frame of CONFLICT, since it involves a high degree of aggression, although situated in a ‘controlled’ sport environment.


Stimulus 4 (took the lead) which, quite surprisingly, did not show any activation in terms of affect, and was rated as neutral both in terms of valence and arousal, was still described by the adjectives happy,

motivated, proud, and supportive. This again reflects the importance of background knowledge structures in the process of emotion-construction.

In sum, the joint results from Questionnaires 3 and 4 suggest that all of the stimuli, i.e. metaphorical expressions corresponding to the conceptual key POLITICS IS SPORT managed to provoke a certain degree of an emotional response with the experimental subjects. This in turn supports the theoretical predictions that conceptual metaphors in the context of political discourse can provoke an emotional response with the audience, i.e. readers, which further serves as a potent mechanism of persuasion.

General Discussion 131

5. General Discussion Following the introduction of the main tenets of the theoretical framework used in the present paper,

the two main parts of the present research along with the main results and implications of these results were presented. The current section will give a comprehensive overview of the main aims of the present research, and the obtained results. Additionally, the potential for interaction between the three main models that constitute the theoretical framework (see section 2 for details), and the implications for future research will also be addressed. Furthermore, the forthcoming paragraphs will also provide a thorough elaboration of the main research questions introduced in section 4.

5.1. Research Questions 1 through 6

As discussed in section 4, the first six research questions were addressed in the first part of the present research. These research questions and the overview of the main results pertaining to each of them are presented below.

[1] What are the basic functions of conceptual metaphors in the political discourse of daily

newspapers?

Based on the findings presented in section 4.1, it can be concluded that the basic function of conceptual metaphors in the political discourse of daily newspapers is that of persuasion, which is aligned with the results presented in previous research (e.g. Charteris-Black, 2004; 2009; 2011; Mio, 1996, 1997; Bougher, 2012; Coulson, 2006a; Coulson & Oakley, 2006). This function is reflected in the fact that metaphors facilitate the installment of perception management mechanisms, which are devised to both direct and constrain the attention of the electorate. The specific mechanisms that facilitate this process are addressed in more detail below.

[2] How are these functions achieved, i.e. which mechanisms are they licensed by (for example, are there mechanisms like emotional appeal, highlighting and hiding, and simplification)?

The present research has shown that the basic mechanisms that facilitate the persuasive function of

metaphors include simplification, marginalization (i.e. highlighting and hiding), and the ability to provoke an

emotional response with the electorate. For instance, SPORT metaphors are especially suitable for simplification, owing to the fact that they reduce the complexity of the political process to a sport event, which in turn causes the emotional involvement of the audience. A good example of highlighting and hiding can be found in CONFLICT metaphors, where, for example, depicting the political process as a battle foregrounds the immediate threat, and places emphasis on the construed foe, while it completely backgrounds more specific


elements of the political process. This also reflects the idea of “polarization,” as discussed in Harder (2005). Still, it needs to be stressed that both CONFLICT and SPORT metaphors are often used to both simplify more complex issues, and to place emphasis on only certain aspects of an ongoing issue. When it comes to the potential for provoking an emotional response, the corpus analysis has shown that both CONFLICT and SPORT metaphors exploit this mechanism. In addition to background knowledge recruitment, emotional reactions can also be provoked by the backward projections, and the coupled elaboration of the blend. With the presented results in mind, it is obvious that the current paper managed to corroborate the findings from previous research (e.g. Mio, 1997; Charteris-Black, 2009; 2011; Ferrari, 2007).

[3] Which groups of metaphors (based on their conceptual keys) are dominant, and are the findings consistent with the results from previous research?

A descriptive statistical overview of the corpus revealed the dominance of CONFLICT metaphors, which

took up 59.58% of the corpus, while SPORT metaphors occupied 40.42% of the corpus. The dominance of these two conceptual keys reflects the competitive nature on which the mainstream Western culture is grounded. In other words, the dominance of CONFLICT metaphors stresses “the extent to which the frame of conflict is entrenched not only in political news reporting, but in the dominant social and political paradigms of Western society” (Burnes, 2011: 2165). Additionally, the immediate context of primary elections for the presidential nomination that were under way in the Republican Party in December 2011 can also be used to justify such frequency of the two conceptual keys. The obtained results are in line with some reports from the previous research in the field (e.g. Burnes, 2011; Semino & Masci, 1996; Howe, 1988). Note, however, that “Koller and Semino (2009: 28) observed a relatively low frequency of war and sport metaphors in the context of German elections, as did Wei (2003: 126) in the case of Taiwan” (cited in Burnes, 2011: 2167). Additionally, Radić-Bojanić and Silaški (2008) suggest that “it is necessary to point to the obvious absence of sport metaphors in post-election periods” (151), meaning that both the cultural component and the immediate socio-political context will play an essential role in determining the use and frequency of specific conceptual keys.

[4] What type of networks is dominant: single-scope, or double-scope?

Based on the results of corpus analysis which was concerned with the investigation of conceptual metaphors of CONFLICT and SPORT used in the political discourse of daily newspapers, it can be concluded that single-scope networks showed absolute dominance, since no instances of double-scope networks have been recorded. This can be explained by the highly entrenched nature of the two conceptual keys. Namely, these entrenched mechanisms serve as strong constraints for the choice of input spaces, as well as for their topological prealignment. Furthermore, as shown in section 4.1, the organizing frames of the inputs usually constitute pre-built compressions available for recruitment. With that in mind, it is obvious why the conceptual integration networks from the present corpus are not expected to show any serious conceptual clashes at the


level of the organizing frames of the respective input spaces. The existing conceptual frame-level clashes are resolved in a relatively simple manner, by projecting the organizing frame of the source input as the organizing frame of the blend. Additionally, it has been suggested in previous research that conventional metaphors are more suitable for use in the political discourse owing precisely to their highly entrenched nature (e.g. Mio, 1997). As a result, they can go on undetected, and perform their persuasive function undisturbed. A novel metaphor would be more likely to appear in the form of a double-scope network, owing to a more elaborate conceptual clash at the frame level. This could in turn cause it to draw more attention, thereby constraining its persuasive potential. In summary, the obtained results pertaining to network structure are plausible, and seem to facilitate the persuasive function of conceptual metaphors in the political discourse of daily newspapers.

[5] Which vital relations are dominant in metaphorical networks, and what happens to them during compression?

The most dominant outer space vital relation recorded in the present corpus was the vital relation of

Analogy that existed between the inputs in metaphorical conceptual integration networks. Such findings suggest that the present corpus consists of single-scope networks in which inputs are not contained within a larger history, i.e. there is no obvious relevance of one input space for the other (Fauconnier & Turner, 2002). In addition to the vital relation of Analogy, there was also the outer space vital relation of Disanalogy that usually appeared between the means for achieving the goals in the respective inputs. Both Analogy and Disanalogy were compressed into inner-space vital relations of Uniqueness in the blend. Furthermore, the compression of the outer-space vital relation of Disanalogy into the inner-space vital relation of Uniqueness in the blend led to the development of the central emergent structure in each network, respectively (see section 4.1 for details).

[6] Is there any similarity between the forms of emergent structures that appear in the present corpus?

A detailed corpus analysis showed a high degree of systematicity in the forms of emergent structures in both CONFLICT and SPORT metaphors, and an overview of these structures is presented in Table 14. Such systematicity is explained by the highly entrenched nature of the two conceptual keys. Additionally, it seems that the systematic nature of compression (in the sense of Fauconnier & Turner, 2008) leads to the systematicity of emergent structures. For more information, consult subsection 4.1.2.1.2 where this issue has already been discussed in great detail.


CONFLICT metaphors Form of emergent structure Number of occurrences in the corpus Percentage

A harms B 126 88.11%

A prevents B from being harmed 15 10.49%

Other 2 1.40%

SPORT metaphors Form of emergent structure Number of occurrences in the corpus Percentage

A improves the position of B 67 69.07%

A harms B 11 11.34% A and B are united by a common goal 9 9.28%

Other 10 10.31%

Table 14. Emergent structures in CONFLICT and SPORT metaphors

5.2. Research Questions 7 and 8

[7] Can conceptual metaphors from the political discourse of daily newspapers provoke an actual, tangible emotional response with the audience, i.e. readers?

[8] How was that emotional response depicted in terms of affect and more specific emotion concepts?

The final two research questions were addressed in the second part of the present research. Namely, based on the methodology introduced in section 4.2, the second part of the present research was designed to determine whether there were any actual emotional responses to the stimuli presented in type-one and type-two questionnaires. The obtained results showed that all stimuli were able to produce a certain degree of an emotional reaction with the experimental subjects. Namely, the emotional reaction was measured based both on the affective response in terms of valence and arousal, and on more specific emotion concepts that were presented in the form of 4 positive and 4 negative emotion concepts.

In terms of affect, the dominant response was recorded on the valence dimension, and the obtained results showed valences ranging from unpleasant to neutral. Only one stimulus, team, showed a positive valence. As far as arousal is concerned, the values of arousal ranged from moderately non-arousing to neutral, and there were no instances of positive arousals. The fact that the two type-one questionnaires recorded the dominant affective response in terms of valence, whereas arousal appeared to be a little harder to grasp, is consistent with the discussions presented in Barrett (2006), and Duncan and Barrett (2007), where the authors suggest that the self-report measurements of affect capture the valence dimension more easily.


Type-two questionnaires also showed consistency in results. The dominant emotion concepts for describing the stimuli corresponding to the conceptual key of CONFLICT were motivated, supportive, and alarmed. On the other hand, the dominant emotion concepts for describing the stimuli corresponding to the conceptual key of SPORT were happy, motivated, supportive, and alarmed. Such findings show a high degree of similarity in the descriptions of emotional reactions based on specific emotion concepts.

As discussed in subsections 4.2.3 and 4.2.6, responses described, for example, by the feelings of alarm can be directly linked to the negative valence. Responses like motivation and support, on the other hand, do not appear to arise directly from the negative affect, but are rather the products of background knowledge structures activated during the processes of meaning and emotion construction. Such findings are in line with the basic tenets of the Conceptual Act Theory of Emotion, according to which the two main components of emotional experience are the categorization of core affect, and the conceptual knowledge about emotions (Barrett, 2006; Lindquist & Barrett, 2008). This in turn stresses the importance of context both in the process of meaning construction, and emotion construction.

Additionally, it is worth stressing that the only stimulus rated with positive valence was the team

metaphor, from the group of SPORT metaphors. This stimulus was also rated by neutral arousal, and the response in terms of specific emotion concepts was described by adjectives happy, motivated, proud, and supportive. Apart from being directly related to positive valence, the choice of these adjectives can also be reinforced by the recruited background knowledge structures related to the organizing frame of TEAM. Furthermore, such reaction can also be justified by the generalized form of the emergent structure, A and B are united by a common goal, typical of TEAM metaphors, which can, via background projections and the coupled elaboration of the blend, serve to further constrain or even amplify the recorded emotional reaction. In brief, it is completely plausible that describing a political group as a team provoked feelings of happiness, motivation, pride, and support, where these can be understood as joint products of positive valence, and the recruited background knowledge structures from the TEAM frame.

5.3. Parts 1 and 2 of the Present Research – Points of Intersection The main aim of the first part of the present research was to investigate the structure of

metaphorical conceptual integration networks, and to explore the mechanisms that these networks might utilize to achieve their persuasive function. This part of the research managed to predict that one of the dominant mechanisms of persuasion resides in metaphors’ ability to provoke an emotional reaction, which is in line with the results of the previous research in the field conducted in the CMT framework (e.g. Mio, 1996;1997; Charteris-Black, 2009; 2011; etc.). The main aim of the second part of the present research was to test whether conceptual metaphors in political discourse can provoke an actual emotional response with the audience. With that in mind, the second part of the research poses as a practical tool for testing


the theoretical prediction of the first part of the research. Based on the results presented so far, it can be concluded that the theoretical prediction pertaining to the metaphors’ potential to resonate with the audience at the emotional level derived from the first part of the research was successfully confirmed in the second stage.

In addition to this obvious point of intersection, further connections between the two parts of the present research can be established. Namely, the Conceptual Act Theory of Emotion stipulates that the two main elements in the process of emotion construction are core affect, and the conceptual knowledge about emotions (e.g. Barrett, 2006; Lindquist and Barrett, 2008). Furthermore, the present research has shown that an important factor in the process of emotion construction is also imposed by the meaning construction network, more precisely the background knowledge structures associated to the organizing frames of the input spaces, and the organizing frame of the blend. Additionally, once the blend is constructed, it can further direct and constrain the process of emotion construction via backward projections, and coupled elaboration. Bearing in mind that backward projections, recruitment of additional background knowledge structures used to complete the emergent structure, and the coupled elaboration of the blend are directly linked to the development of emergent structure in the network, further connections can be established between the developed emergent structures, and the recorded emotional reactions.

For instance, the stimulus barrage of attacks54 developed the emergent structure in the generalized form A harms B. In turn, this stimulus was rated by negative valence (2.41; p=.000), and neutral arousal (4.07; p=.091), and by the adjectives angry and alarmed. Such affect categorization can be attributed to the quality of the stimulus, and the background knowledge structures associated to the ATTACK frame. Additionally, the negative valence can be caused by the ELECTION frame that serves as the organizing frame of the target input. Namely, the context of politics alone can induce the feeling of unpleasantness. The feelings of anger and alarm can be linked directly to the negative valence, but can also be interpreted in terms of the background knowledge structures introduced by the organizing frames of the two inputs. Additionally, the developed emergent structure can yield backward projections to the input spaces, and along with the coupled elaboration of the blend, serve to influence both the initial affective state, and the recruitment of specific emotion concepts. An additional factor that must not be neglected is the influence of the immediate discourse context.

The stimulus showdown55 also developed the generalized emergent structure in the form of A harms B, and it was rated by negative valence (3.88; p=.009), neutral arousal (4.65; p=.525), and the adjectives motivated and supportive. Both the negative valence, and the positive feelings of motivation and support constitute plausible results. However, while the negative valence depicts the initial unpleasant reactions to the concept ‘showdown’ introduced by the organizing frame of the source input, feelings of motivation and support

54 The entire sentence used in Questionnaires 1 and 2: “As the perceived frontrunner, he is drawing a barrage of attacks from his rivals.” 55 The entire sentence used in Questionnaires 1 and 2: “The political party began the preparations for the showdown with its opponents.”


can be accounted for by the immediate discourse context, where it can be argued that the reader feels support for the political party preparing for a showdown, and is motivated by this idea. In that sense, in addition to the immediate context, an important variable is also the background knowledge activated during the construction of the blend, as well as the possible backward projections and coupled elaboration, both of which stem directly from the blend. Note that the stimulus re-election fight56 was also described in the same terms as the stimulus showdown, the only difference being that the arousal in this case was negative instead of neutral. Such similarity in reactions to the two stimuli can be accounted for by the same form of emergent structure, as well as by the high level of relatedness between the frames of SHOWDOWN and FIGHT that pose as the organizing frames of the two respective source inputs.

The stimulus (war) front57 was rated by negative-to-neutral valence (3.99; p=.026), neutral arousal (4.08; p=.065), and described by the adjective supportive. This stimulus also developed the emergent structure in the form A harms B. Negative-to-neutral valence can be described by the quality of the stimulus and by the immediate context, as well as by the (WAR) FRONT frame that serves as the organizing frame of the source input. Additionally, negative-to-neutral valence can also be caused by the POLITICAL PROCESS frame that is the organizing frame of the target input, where the context of politics alone is also likely to produce unpleasant feelings. The adjective supportive can be accounted for by the background knowledge structures recruited in the process of meaning construction, where the reader feels supportive of those on the war front. Additionally, the developed emergent structure can further constrain, or even amplify the recorded emotional response via its backward projections and coupled elaboration.

The stimulus sparring58 from the group of SPORT metaphors has transparent links to the CONFLICT frame. This fact explains that the results obtained from Questionnaires 3 and 4 are similar to those obtained from the stimuli belonging to the group of CONFLICT metaphors. Namely, sparring was described by negative valence (3.09; p=.000), neutral arousal (4.14; p=.190), and the adjectives supportive and alarmed. This stimulus developed the emergent structure in the generalized form A harms B, as was the case with the majority of CONFLICT metaphors. The recorded negative valence can be accounted for by the quality of the stimulus and the immediate context, as well as by the frame of ELECTIONS which serves as the organizing frame of the target input. The feeling of alarm can be linked both to the negative valence, and to the recruited background knowledge structures, while the feeling of support is inherently linked to the BOXING frame, whereby the reader feels support for his fighter. Again, backward projections and the coupled elaboration of the blend can further influence the process of emotion construction, where, for example, the coupled elaboration of the blend can increase either the feeling of support or alarm.

56 The entire sentence used in Questionnaires 1 and 2: “Facing a tough re-election fight in five months, the president is presenting himself as a man of experience.” 57 The entire sentence used in Questionnaires 1 and 2: “I would like the president to be a bit more courageous on this front, about the labor market.” 58 The entire sentence used in Questionnaires 3 and 4: “The day began with the two candidates sparring with each other.”


The stimulus break out of the pack59 developed the emergent structure in the form A improves the position of B. It was described by negative valence (3.09; p=.000), neutral arousal (4.67; p=.498), and the adjectives happy, motivated, proud, supportive, and alarmed. The negative valence can be accounted for by the immediate context of political discourse, as well as by the PRESIDENTIAL ELECTIONS frame that organizes the target input. The feelings of happiness, motivation, pride, and support can be linked to the RACE frame, where the reader experiences these feelings when one of the runners takes the lead. The feeling of alarm is directly linked to the negative valence, and can also be accounted for by the context of politics imposed by the discourse, and the organizing frame of the target input, i.e. the PRESIDENTIAL ELECTIONS frame. In that sense, the reader could also be alarmed because an unwanted candidate is taking the lead. Additionally, both backward projections and coupled elaboration can serve to influence the shape and amplitude of the emotional response.

The stimulus last minute sprints60 was described by negative valence (3.37; p=.000), neutral arousal (4.56; p=.828), and the adjectives motivated and supportive. The present stimulus developed the emergent structure in the form A improves the position of B. Again, like in the previous case, the negative valence results from the immediate context of elections, while the feelings of motivation and support stem from the background knowledge structures associated with the RACE frame. Coupled elaboration of the emergent structure, as well as possible backward projections can serve to amplify the recorded feelings.

The stimulus took the lead61 developed the emergent structure in the form A improves the position of B, and the emotional response to it was recorded by neutral valence (4.54; p=.869), neutral arousal (4.77; p=.342), and the adjectives happy, motivated, proud, and supportive. The neutral affect can be accounted for by the highly entrenched nature of the stimulus, in the sense that it is a worn-off expression likely to be taken for granted. The recorded feelings of happiness, motivation, pride, and support do not seem to go hand in hand with the recorded affect. However, in this case the process of emotion construction appears to be dominated by the conceptual knowledge about emotions, and the background knowledge structures introduced by the RACE frame, while the categorized affect remains neutral. With that in mind, the coupled elaboration of the emergent structure along with the possible backward projections can play an important part in amplifying the experienced emotions.

Quite surprisingly, the stimulus presidential race 62 did not provoke any positive feelings that are expected to appear in relation to the RACE frame. Based on the other results obtained from Questionnaires 3 and 4, it can be concluded that the dominant role in provoking the emotional reaction with this particular stimulus was played by the immediate context of presidential elections, also introduced by the organizing frame of the target input. Although the given stimulus developed the emergent structure in the form A 59 The entire sentence used in Questionnaires 3 and 4: “One of the presidential candidates attempted to break out of the pack.” 60 The entire sentence used in Questionnaires 3 and 4: “The two candidates continued their last-minute sprints before the elections.” 61 The entire sentence used in Questionnaires 3 and 4: “One of the presidential candidates took the lead.” 62 The presidential race still remains remarkably unsettled.


improves the position of B, the reaction in terms of emotion concepts was recorded by the adjective alarmed, while both valence (2.94; p=.000) and arousal (3.67; p=.005) showed negative values. In that sense, the feeling of alarm can be directly linked to the negative affect. Therefore, unlike the previous case in which affect categorization did not contribute much to the process of emotion construction, and where that process was governed mostly by the background knowledge structures, here, affect categorization appears to place a very strong influence on this process. Additionally, the feeling of alarm can also be linked to the background knowledge associated to the organizing frame of the target input, i.e. the PRESIDENTIAL ELECTIONS frame. Consequently, the emotional response can be further influenced by backward projections to the target input space.

Note that similar analyses to those presented in the previous paragraphs can be applied to the rest of the stimuli from the questionnaires used in the second part of the present research. In summary, it can be concluded that the results obtained from the two parts of the present research are not only comparable, but also provide a valuable joint account that sheds light on the links between the processes of emotion construction and meaning construction, in terms of how the two interact to produce a fully developed human scale experience. Namely, the background knowledge structures introduced by the organizing frames of the input spaces influence both the process of meaning and emotion construction. Furthermore, the developed emergent structure can serve as an additional constraint that works via backward projections and coupled elaboration. Bearing in mind that most words have an inherent affective quality (Duncan & Barrett, 2007), the influence of affect and emotion construction on the process of meaning construction should not be neglected either. Once again, such interaction is licensed by the fact that the specific brain circuitries that support affect are largely shared by the circuitries that support cognitive processes (ibid.)

5.4. Additional Remarks

The present research also offers some conclusions pertaining to the theoretical framework introduced in section 2.2.8, in relation to the joint use of the CMT and CBT models. Namely, based on the results of the corpus analysis obtained in section 4.1, it is obvious that the CBT framework provides a more thorough insight into the structure and mechanisms at work in metaphorical single-scope conceptual integration networks. Additionally, the possibility of backward projections and coupled elaboration of the blend present themselves as potentially valuable variables in the process of emotion construction. The CMT model, on the other hand, resides in a lot more restricted framework that sanctions only unidirectional projections from the source to the target domain. In light of the present research, such an approach is dubbed insufficient, and unable to provide a complete account of the process of meaning construction.

Additionally, it is worth noting that the Invariance Principle (Lakoff, 1990; 2006[1993]), which, apart from the initial choice of source and target domains, presents itself as perhaps the most important constraint of


the CMT framework, remains almost completely unutilized in the previous research in the field. Namely, few authors have bothered to include it in their analyses. Furthermore, the CMT model gives a much more detailed account pertaining to the sociological function of metaphors in political discourse, emphasizing the notions of highlighting and hiding, and directing attention. However, while these processes are intuitively plausible and obvious, the exact cognitive mechanisms that underlie them remain unrevealed by this theoretical model, in the sense that the CMT neglects the potential interaction between the organizing frames of the source and target domains, as well as more serious clashes between them. Furthermore, the notions of compression and emergent structure are also omitted, and the model seems to satisfy itself at the level of metaphorical entailments. An additional point that the CMT model neglects is the need for the blended space, because the cross-domain transfer, although essential, does not finalize the process of meaning construction. Analogies and similarities established in cross-domain mappings simply need to be projected somewhere so that they can be viewed as a single element. Leaving a conceptual metaphor as a mere two-domain structure would require that two elements (one element from each domain) occupy the same functional slot during the cognitive process, which would in turn cause too much cognitive tension and render the system unstable. In short, for a conceptual metaphor to work, the system that supports it needs to be fully optimized, and in the state of an equilibrium. Otherwise it will disintegrate.

However, despite its obvious shortcomings, the CMT can still be considered a borderline case of the CBT framework (Fauconnier & Turner, 1994; 1995; Grady, Oakley, & Coulson, 2007[1999]). In addition to the similarities between the two theoretical models outlined in section 2.2.8, which serve to reinforce the discussion in favor of their joint use, the present research will present another argument in order to further substantiate such an approach. Namely, while the CBT model offers obvious advantages for practical analyses, the CMT framework remains indispensable for the purposes of corpus construction. The use of the CBT model in this process would be too time-consuming and impractical, since corpus construction requires the identification of specific conceptualizations, and the conceptual keys that they belong to. This is achieved by identifying metaphor keywords or key expressions, which are then grouped according to the conceptual metaphors and keys they instantiate. Therefore, the CMT framework alone satisfies the needs of corpus construction. Issues like, for instance, the form of emergent structure, are not necessary at this point, and, as already suggested, would in fact halt the entire process. In brief, the value of the CMT framework in corpus construction further reinforces the idea of its concomitant use with the CBT model. Such an approach will facilitate both the process of corpus construction, and yield more elaborate and systematic results when it comes to metaphor analysis. Additionally, apart from the nature of cognitive processes that underlie metaphor as a phenomenon, CMT can serve to provide a further account of its sociocultural dimension.


5.5. Section Summary The current section discussed the most important results obtained from the present research. Namely,

the initially introduced research questions have been addressed in great detail, which was followed by the establishment of a direct connection between the results obtained in the two parts of the present research. Finally, the section closed by discussing the implications for the joint use of the CBT and CMT frameworks in light of the present results.

Conclusion 142

6. Conclusions and Suggestions for Further Research

In summary, the present research gave way to some very interesting, and potentially relevant findings. First of all, it was established that the entrenched nature of conceptual keys of CONFLICT and SPORT metaphors led to a high systematicity in the forms of emergent structures (see Table 14 for details). Additionally, the systematic nature of compression (in the sense of Fauconnier & Turner, 2008) yielded the elaborate “molecule-model” that depicts both the on-line process of meaning construction, and the recruitment of background knowledge structures, owing to the fact that inputs are often themselves pre-built compressions available for recruitment. The present research also discussed the persuasive function of metaphors in political discourse and managed to corroborate the findings from previous research. Additionally, one of the main mechanisms of persuasion, the ability to provoke an emotional reaction, was tested empirically, and the obtained results confirmed the initial theoretically derived predictions. Furthermore, certain points of intersection between the Conceptual Act theory of emotion and the CBT have also been outlined. The obtained results suggest that for a fully developed, first-person, human scale experience to take place, both the semantic and the emotional dimension need to be accounted for. Otherwise, the constructed experience will be incomplete. However, the exact nature of the interplay between the processes of meaning and emotion construction remains to be addressed in further research.

In addition to the findings mentioned in the previous paragraph, the paper was also concerned with the notions of compression, vital relations, the generic space problem, backward projections, coupled elaboration, and some other issues relevant in the present context. However, bearing in mind that all of those findings have already been discussed in great detail in the section 5, in order to avoid a reiteration of the already introduced conclusions, the remainder of the present section will provide some suggestions for further research that can both improve and complete the research conducted in the present paper.

The first obvious pointer for further research is situated in the fact that type-two questionnaires could be improved by the factor analytic methodological approach. Namely, this would require a separate research, where a longer list of emotion concepts would be reduced to a shorter and a more reliable one. The aim of such an approach would be to develop a highly reliable and specialized emotion concepts checklist, designed to rate the responses to various stimuli from the context of political discourse, including metaphorical expressions. Another valuable parameter to be taken into account includes cultural differences. Namely, while a large sample of, for example, Serbian participants, would yield a reliable checklist for the context of Serbian culture, that same list may not be as effective in the context of German or Polish culture. Therefore, the final goal of such line of research should be the development of an internationally reliable checklist that would in turn facilitate the comparison of research findings across various cultures and various socio-political contexts.

A line of research that utilizes self-reports could also be combined with a priming study. For example, the experimental subjects could be divided into three groups, where the first group would not be primed, the

Conclusion 143

second group would be primed for negative affect by listening to a “stressful” song, while the third group could be primed for positive affect by listening to a “relaxing” song. Such priming would constitute a part of the immediate context, and further statistical analyses of the obtained results could serve to determine to what extent the priming procedure influenced the participants’ reactions. This would in turn reveal the importance of the immediate context in the construction of emotional experience.

Additionally, further research could also benefit from a more aggressive, and a more hi-tech interdisciplinary approach that would include the measurements of the N400 evoked potentials during exposure to metaphors extracted from the political discourse. Such measurements, coupled with well though out EEG or fMRI studies could serve to further explore the specific neural circuitries in charge of metaphorical language processing. In the context of emotional reaction measurements, in addition to the self-report measurements conducted via appropriate questionnaires, the research could also include a set of neuro-physiological measurements, following the methodological guidelines presented in Nešić et al. (2009; 2010). This way, the results obtained from self-reports could be cross-referenced against a set of more objective recordings of somatosensory reactions.

Future research in the domain of the CBT should also attempt to move in the direction of a firmer empirical grounding. Namely, although there are circumstantial evidence that, for example, support the notion of highly specified frames in the form of situated conceptualizations (Barsalou, 2009), the greater part of the theoretical framework still continues to hover over the phenomena that it describes, and the concepts that it utilizes. Apart from the more exotic laboratory approaches described in the previous paragraph, perhaps a low-tech alternative in the form of simple questionnaires might also do the trick.

* * *

In conclusion, the CBT framework poses as a very powerful tool that can shed more light onto the mysterious processes that take place in the human mind during meaning and emotion construction in the context of conceptual metaphors. The results presented in this research reveal an intrinsic link between the emotional and semantic meaning, thereby suggesting that both dimensions are required in order for the complete human scale experience to be built. Further research in the field, as well as future interdisciplinary endeavors should provide additional insights into these processes. In the meantime, the blending paradigm continues its evolutionary path.

References 144

References [1] Antović, M. (2010). From oceanic feeling to image schemata – embodied mind and the construction of

identity. In Identity Issues – Literary and Linguistic Landscapes, Lopičić, V. and B. Mišić Ilić, eds., 177-194. Cambridge Scholars Publishing.

[2] Antović, M. (2013). Musical Metaphor Revisited: Primitives, Universals and Conceptual Blending. In review.

[3] Antović, M. and Tasić, M. (2011). Muzika na granici i teorija konceptualne integracije: analiza jedne pesme sa juga Srbije. Teme, 4, 1521-1542.

[4] Barlett, C. P. and Gentile, D. A. (2011). Affective and emotional consequences of the mass media. In The Routledge Handbook of Emotions and Mass Media, Döveling, K., C. Von Scheve and E. A. Konijn, eds., 61-78. Oxon/New York: Routledge.

[5] Barsalou, L. W. (2009). Simulation, situated conceptualization, and prediction. Philosophical Transactions of the Royal Society, B: Biological Sciences, 364(1521), 1281-1289.

[6] Barrett, L. F. (2006). Solving the Emotion Paradox: Categorization and the Experience of Emotion. Personality and Social Psychology Review, 10(1), 20-46.

[7] Bougher, L. D. (2012). The Case for Metaphor in Political Reasoning and Cognition. Political Psychology, 33(1), 145-163.

[8] Burnes, S. (2011). Metaphors in press reports of elections: Obama walked on water, but Musharaf was beaten by a knockout. Journal of Pragmatics, 43(8), 2160-2175.

[9] Charteris-Black, J. (2004). Corpus Approaches to Critical Metaphor Analysis. Basingstoke and New York: Palgrave Macmillan.

[10] Charteris-Black, J. (2009). Metaphor and Political Communication. In Metaphor and Discourse, Musolff, A. and J. Zinken, eds., 97-115. New York: Palgrave Macmillan.

[11] Charteris-Black, J. (2011). Politicians and Rhetoric: The Persuasive Power of Metaphor, 2nd Edition. New York: Palgrave Macmillan.

[12] Cienki, A. (2007). Frames, Idealized Cognitive Models, and Domains. In The Oxford Handbook of Cognitive linguistics, Geeraerts, D. and H. Cyckens, eds., 170-187. Oxford: Oxford University Press.

[13] Coulson, S. (2001). Semantic Leaps: Frame-Shifting and Conceptual Blending in Meaning Construction. Cambridge: Cambridge University Press.

[14] Coulson, S. (2006a). Conceptual blending in thought, rhetoric, and ideology. In Cognitive Linguistics: Current Applications and Future Perspectives, Kristiansen, G., M. Achard, R. Dirven and F. J. I. Ruiz de Mendoza, eds., 187-208. Berlin/New York: Mouton de Gruyter.

[15] Coulson, S. (2006b). Metaphor and Conceptual Blending. In Encyclopedia of Language and Linguistics, 2nd Edition, Keith, B., Editor-in-Chief, 32-39. Oxford: Elsevier.

[16] Coulson, S. and Oakley, T. (2000). Blending Basics. Cognitive Linguistics, 11-3/4, 175-196.

[17] Coulson, S. and Oakley, T. (2005). Blending and Coded Meaning: Literal and figurative meaning in cognitive semantics. Journal of Pragmatics, 37(10), 1510-1536.

References 145

[18] Coulson, S. and Oakley, T. (2006). Purple Persuasion: Conceptual Blending and Deliberative Rhetoric. In Cognitive Linguistics: Investigations across languages, fields, and philosophical boundaries, J. Luchenbroers, ed., 47-65. Amsterdam: John H. Benjamins.

[19] Coulson, S. and Oakley, T. (2008). Connecting the dots. Mental spaces and metaphoric language in discourse. In Mental Spaces in Discourse and Interaction, Oakley, T. and A. Hougard, eds., 27-50. Amsterdam/Philadelphia: John Benjamins Publishing Company.

[20] Crawford, J., R. and Henry, J. D. (2004). The Positive and Negative Affect Schedule (PANAS): Construct validity, measurement properties and normative data in a large non-clinical sample. British Journal of Clinical Psychology, 43(3), 245-265.

[21] Cronk, B. C. (2006). How to Use SPSS: A Step-by-Step Guide to Analysis and Interpretation, 4th Edition. Glendale: Pyrczak Publishing.

[22] Dancey, C. P. and Reidy, J. (2007). Statistics Without Maths for Psychology, 4th Edition. Harlow: Pearson Education Limited.

[23] Drulák, P. (2006). Motion, Container and Equilibrium: Metaphors in the Discourse about European Integration. European Journal of International Relations,12(4), 499-531.

[24] Duncan, S., and Barrett, L. F. (2007). Affect is a form of cognition: A neurobiological analysis. Cognition and Emotion, 21(6), 1184-1211.

[25] Evans V. and Green M. (2006). Cognitive Linguistics: An Introduction. Edinburgh: Edinburgh University Press Ltd.

[26] Fauconnier, G. (1994). Mental Spaces: Aspects of Meaning Construction in Natural Language. Cambridge: Cambridge University Press.

[27] Fauconnier, G. (1997). Mappings in Thought and Language. Cambridge: Cambridge University Press.

[28] Fauconnier, G. (2007). Mental Spaces. In The Oxford Handbook of Cognitive linguistics, Geeraerts, D. and H. Cyckens, eds., 351-376. Oxford: Oxford University Press.

[29] Fauconnier, G. (2008). How Compression Gives Rise to Metaphor and Metonymy. A talk given at the 9th Conference on Conceptual Structure, Discourse, and Language (CSDL9), at the University of California, San Diego (UCSD) - Department of Cognitive Science. Retrieved from: https:// www. youtube. com/ watch?v=kiHw3N6d1Js, on 8 February 2013, 18:52.

[30] Fauconnier, G. and Turner, M. (1994). Conceptual Projection and Middle Spaces. UCSD Cognitive Science Technical Report 9401. Retrieved from: http://www.cogsci.ucsd.edu/ research/documents/ technical/9401.pdf, on 18/11/2012, 15:48.

[31] Fauconnier, G. and Turner, M. (2000). Compression and global insight. Cognitive Linguistics, 11-3/4, 283-304.

[32] Fauconnier, G. and Turner, M. (2002). The Way We Think: Conceptual Blending and Mind’s Hidden Complexities. Basic Books, New York.

[33] Fauconnier, G. and Turner, M. (2006[1998]). Conceptual Integration Networks. In Cognitive Linguistics: Basic Readings (Cognitive Linguistics Research 34), Geeraerts, D., ed., 303-373. Berlin/New York: Mouton de Gruyter.

[34] Fauconnier, G. and Turner, M. (2008). Rethinking Metaphor. In The Cambridge Handbook of Metaphor and Thought, Gibbs, R. W. Jr., ed., 53-66. Cambridge: Cambridge University Press.

References 146

[35] Ferrari, F. (2007). Metaphor at work in the analysis of political discourse: investigating a ‘preventive war’ persuasion strategy. Discourse & Society, 18(5), 603-625.

[36] Fillmore, C., J. (2006[1982]). Frame Semantics. In Cognitive Linguistics: Basic Readings (Cognitive Linguistics Research 34), Geeraerts, D., ed., 373-400. Berlin/New York: Mouton de Gruyter.

[37] Gibbs, R. W. Jr. and Colston, H. L. (2006[1995]). The cognitive psychological reality of image schemas and their transformations. In Cognitive Linguistics: Basic Readings (Cognitive Linguistics Research 34), Geeraerts, D., ed., 239-268. Berlin/New York: Mouton de Gruyter.

[38] Grady, J. E. (2005). Primary metaphors as inputs to conceptual integration. Journal of Pragmatics, 37(10), 1595–1614.

[39] Grady, J. E. (2007). Metaphor. In The Oxford Handbook of Cognitive Linguistics, Geeraerts, D. and H. Cyckens, eds., 188-213. Oxford: Oxford University Press.

[40] Grady, J. E., Oakley, T., and Coulson, S. (2007[1999]). Blending and Metaphor. In The Cognitive Linguistics Reader, Evans, V., B. K. Bergen and J. Zinken, eds., 420-440. London/Oakville: Equinox Publishing Ltd.

[41] Gray, E., K. and Watson, D. (2007). Assessing positive and negative affect via self-report. In Handbook of Emotion Elicitation and Assessment, Coan, J. A. and J.J.B. Allen, eds., 171-183. Oxford: Oxford University Press.

[42] Harder, P. (2005). Blending and polarization: Cognition under pressure. Journal of Pragmatics, 37(10), 1636-1652.

[43] Howe, N. (1988). Metaphor in Contemporary American Political Discourse. Metaphor and Symbolic Activity, 3(2), 87-104.

[44] Hutchins, E. (2005). Material anchors for conceptual blends. Journal of Pragmatics, 37(10), 1555–1577.

[45] Joy, A., Sherry, J. F. Jr., and Deschenes, J. (2009). Conceptual blending in advertising. Journal of Business Research, 62(1), 39-49.

[46] Kövecses, Z. (2006). Language, Mind and Culture: A Practical Introduction. Oxford: Oxford University Press.

[47] Kövecses, Z. (2010). Metaphor: A Practical Introduction, 2nd Edition. Oxford/New York: Oxford University Press.

[48] Lakoff. G. (1990). The Invariance Hypothesis: Is abstract reason based on image schemas? Cognitive Linguistics, 1(1), 39-74.

[49] Lakoff, G. (1991). Metaphor and War: The Metaphor System Used to Justify War in the Gulf. Retrieved from: http://georgelakoff.files.wordpress.com/2011/04/metaphor-and-war-the-metaphor-system-used-to-justify-war-in-the-gulf-lakoff-1991.pdf, on 1/11/2012, 22:15.

[50] Lakoff, G. (2006[1993]). The contemporary theory of metaphor. In Cognitive Linguistics: Basic Readings (Cognitive Linguistics Research 34), Geeraerts, D., ed., 185-238. Berlin/New York: Mouton de Gruyter.

[51] Lakoff, G. and Johnson, M. (1980a). The Metaphorical Structure of the Human Conceptual System. Cognitive Science, 4(2), 195-208.

[52] Lakoff, G. and Johnson, M. (2003[1980b]). Metaphors We Live By. London: The University of Chicago Press.

References 147

[53] Lang, A. and Ewoldsen, D. (2011). The measurement of positive and negative affect in media research. In The Routledge Handbook of Emotions and Mass Media, Döveling, K., C. Von Scheve and E. A. Konijn, eds., 79-98. Oxon/New York: Routledge.

[54] Lindquist, K. A. and Barrett, L. F. (2008). Constructing Emotion: The Experience of Fear as a Conceptual Act. Psychological Science, 19(9), 898-903.

[55] Mandler, J. (1992). How to Build a Baby: II. Conceptual Primitives. Psychological Review, 99(4), 587-604.

[56] Mandler, J. (2012). On the Spatial Foundations of the Conceptual System and Its Enrichment. Cognitive Science, 36(3), 421-451.

[57] Mio, J. S. (1996). Metaphor, Politics, and Persuasion. In Metaphor: Implications and Applications, Mio, J. S, and A. N. Katz, eds. 127-164. Mahwah, New Jersey: Lawrence Erlbaum Associates.

[58] Mio, J. S. (1997). Metaphor and Politics. Metaphor and Symbol, 12(2), 113-133.

[59] Mio, J. S., Riggio, R. E., Levin, S., and Reese, R. (2005). Presidential leadership and charisma: The effects of metaphor. Leadership Quarterly, 16(2), 287-294.

[60] Nešić, M., Ćirić, M., Nešić, Milenović, M., and Branković, S. (2009). Valence, arousal and skin conductance of emotional movie scenes. Physioacta, 3(2), 137-147.

[61] Nešić, M., Ćirić, M., Nešić, V., Filipović, J., and Veličković, M. (2010). Impact of anxiety on valence, arousal, and cardiovascular reactivitz to emotional movie scenes. Bulgarian Journal of Psychology, Issue 1-4, 138-149.

[62] Niedenthal, P. M. (2008). Emotion Concepts. In Handbook of Emotions, Lewis, M., J. M. Haviland-Jones, L. F. Barrett, eds., 587-600. New York/London: The Guilford Press.

[63] Oakley, T. (2007). Image Schemas. In The Oxford Handbook of Cognitive Linguistics, Geeraerts, D. and H. Cyckens, eds., 214-235.Oxford: Oxford University Press.

[64] Oakley, T. (2011). Conceptual Integration. In Handbook of Pragmatics, Östman, J. O. and J. Verschueren, eds., 1-24. John Benjamins Publishing Company. Retrieved from: http://academia.edu/1215868/Conceptual_Integration, on 20 March 2013, 22:55.

[65] Paivio, A. and Walsh, M. (1993). Psychological processes in metaphor comprehension and memory. In Metaphor and Thought, 2nd Edition, Ortony, A., ed., 307-328. Cambridge: Cambridge University Press.

[66] Pallant, J. (2011). SPSS Survival Manual. A step by step guide to data analysis using SPSS, 4th edition. Crow’s Nest, Australia: Allen & Unwin.

[67] Posner, J., Russell, J. A., and Peterson, B. S. (2005). The circumplex model of affect: An integrative approach to affective neuroscience, cognitive development, and psychopathology. Development and Psychopathology, 17(3), 715-734.

[68] Radić-Bojanić, B., and Silaški, N. (2008). Sportizacija političkog diskursa – kako metafore prikrivaju političku stvarnost Srbije. Zbornik Matice srpske za filologiju i lingvistiku, 51, 1(2), 139-155.

[69] Ridout, T. N. and Searles, K. (2011). It’s My Campaign I’ll Cry if I Want to: How and When Campaigns Use Emotional Appeals. Political Psychology, 32(3), 439-458.

[70] Ruiz de Mendoza, F. J., and Peña, S. (2005). Conceptual interaction, cognitive operations and projection spaces. In Cognitive Linguistics: Internal Dynamics and Interdisciplinary Interaction, Ruiz de Mendoza, F. J. and S. Peña, eds., 249-280. Berlin/New York: Mouton de Gruyter.

References 148

[71] Russell, J. A. (1980). A Circumplex Model of Affect. Journal of Personality and Social Psychology, 39(6), 1161-1178.

[72] Russell, J., A., Weiss, A., and Mendelsohn, G., A. (1989). Affect Grid: A Single-Item Scale of Pleasure and Arousal. Journal of Personality and Social Psychology, 57(3), 493-502.

[73] Russell, J. A. (2003). Core Affect and the Psychological Construction of Emotion. Psychological Review, 110(1), 145-172.

[74] Saeed, J. I. (2003). Cognitive Semantics. In Semantics, Saeed, J. I., 342-384. Malden: Blackwell Publishing.

[75] Schemer, C. (2012). Reinforcing Spirals of Negative Affect and Selective Attention to Advertising in a Political Campaign. Communication Research, 39(3), 413-434.

[76] Semino, E. and Masci, M. (1996). Politics is football: Metaphor in the discourse of Silvio Berlusconi in Italy. Discourse & Society, 7(2), 243-269.

[77] Silaški, N., Đurović, T., and Radić-Bojanić, B. (2009). Javni diskurs Srbije – kognitivističko-kritička studija. Beograd: Čugura Print.

[78] Silaški, N. and Radić-Bojanić, B. (2010). Premijer kao trener sportskog tima. In Peti međunarodni interdisciplinarni simpozijum Susret kultura, zbornik radova, Subotić Lj. and I. Živančević-Sekeruš, eds., 625-632. Novi Sad: Univerzitet u Novom Sadu, Filozofski fakultet.

[79] Silaški, N., and Đurović T. (2011). The natural force metaphor in the conceptualisation of the global financial crisis in English and Serbian. Zbornik Matice srpske za filologiju i lingvistiku. LIV, 1, 227-245.

[80] Stamenković, D. (2009). Uloga kognitivne semantike u izučavanju prozne književnosti: konceptualno stapanje u basnama. In Savremena proučavanja jezika i književnosti – Godina I / knjiga I – Zbornik radova sa I naučnog skupa mladih filologa Srbije, Kovačević, M. et al. eds., 425–435. Kragujevac: Filološko-umetnički fakultet.

[81] Steinert, H. (2003). The indispensable metaphor of war: On populist politics and the contradictions of the state’s monopoly of force. Theoretical Criminology, 7(3), 265–291.

[82] Thompson, E. R. (2007). Development and validation of an internationally reliable short-form of the positive and negative affect schedule (PANAS). Journal of Cross-Cultural Psychology, 38(2), 227-242.

[83] Thompson, S. (1996). Politics without Metaphors is Like a Fish without Water. In Metaphor: Implications and Applications, Mio, J. S. and A. N. Katz, eds., 185-201. Mahwah, New Jersey: Lawrence Erlbaum Associates.

[84] Turner, M. (2007). Conceptual Integration. In The Oxford Handbook of Cognitive Linguistics, Geeraerts, D. and H. Cyckens, eds., 377-393. Oxford: Oxford University Press.

[85] Turner, M. and Fauconnier, G. (1995). Conceptual Integration and Formal Expression. Metaphor and Symbolic Activity, 10(3), 183-204.

[86] Vidanović, Đ. (2012). Conceptual Blending and Intentionality. Facta Universitatis, Linguistics and Literature Series, 10(1), 41-52.

[87] Watson, D., Clark, L. A., Tellegen, A. (1988). Development and Validation of Brief Measure of Positive and Negative Affect: The PANAS Scales. Journal of Personality and Social Psychology, 54(6), 1063-1070.

References 149

[88] Watson, D, and Clark, L. A. (1994). The PANAS-X: Manual for the Positive and Negative Affect Schedule – Expanded Form. University of Iowa, Iowa Research Online, Department of Psychology Publications. Retrieved from: http://ir.uiowa.edu/psychology_pubs/11, 2 June 2013, 15:30.

[89] Wilson-Mendenhall, C. D., Barrett, L. F., Simmons, K. W., and Barsalou, L. W. (2011). Grounding emotion in situated conceptualization. Neuropsychologia, 49(5), 1105-1127.

Appendix A 150

Appendix A – Questionnaire 1

Appendix A 151

Appendix B 152

Appendix B – Questionnaire 2

Appendix B 153

Appendix C 154

Appendix C – Questionnaire 3

Appendix C 155

Appendix D 156

Appendix D - Questionnaire 4

Appendix D 157

analysis of conceptual metaphors in the

Documents