Show Less
Open access

Reading Autobiographical Comics: A Framework for Educational Settings


Markus Oppolzer

This book updates reader-response criticism as the foundation of aesthetic reading in the classroom by bringing it in line with cognitive theories in literary studies and linguistics. With the help of Gilles Fauconnier and Mark Turner‘s conceptual integration theory, which shares a surprising number of correspondences with Wolfgang Iser‘s The Act of Reading, it is possible to flesh out the latter‘s model of narrative meaning-making. In turn, this allows for a consistent reader-response approach to the medium of comics and auto/biography as one of its dominant genres. The fragmentation of comics narratives, but also of human lives and identities, requires such a theory that can explain how different perspectives and experiences can be blended into an experiential whole.

Show Summary Details
Open access

3 Cognitive (Literary) Studies

←122 | 123→

3 Cognitive (Literary) Studies

3.1 The Return of the Reader

Wolfgang Iser’s ample use of optical metaphors indicates that he is concerned with the mind’s eye and readers’ construction of meaning as a mental process. This part sets out to highlight the striking similarities between core concepts of reader-response criticism and cognitive approaches to language and literature, a circumstance that has been frequently implied or directly pointed out (cf. e.g. Bredella 2010: 24, 78; Hogan 2003: 160; Keen 2010: xi), but not sufficiently explored. When Andreas Müller-Hartmann calls reader-response criticism a ‘modern’ literary theory (cf. 2007: 197), I could not agree more, but it is in need of a re-evaluation and a re-contextualisation in terms of cognitive (literary) studies.

Cognitive literary studies is a sprawling and bewildering new field of inquiry that is prone to speculation and draws its facts and – more often than not – its inspirations from cognitive sciences across a number of disciplines. David Groome identifies these research areas as follows: experimental cognitive psychology, computer modelling, cognitive neuropsychology and cognitive neuroscience (cf. 2014a: 5). For practical purposes, I mostly rely on a select number of authors, such as Antonio Damasio (2000) in the area of neurophysiology, or the work of cognitive psychologists, such as David S. Miall and Don Kuiken (1994, 2002), whose research is explicitly in the area of reading. In the field of cognitive literary studies some names, such as Alan Palmer (2004, 2010) or Suzanne Keen (2010), feature more prominently than others and the same applies to cognitive linguistics, where I mainly rely on George Lakoff and Mark Johnson’s Metaphors We Live By (2003) next to Gilles Fauconnier and Mark Turner’s The Way We Think (2003). I have to leave out whole disciplines, such as philosophy, with enactivism being the sole exception.

What complicates matters further is the historical development of research. It is essential to draw a “distinction between first-generation and second-generation cognitive science” (Caracciolo 2014: 16), by which Marco Caracciolo means a computational and an embodied model respectively. In essence, this is the question of whether the brain is an independent computer-like machine that is carried around by a living organism, or part of a larger integrated network of systems/organs that spans the entire biological body and transacts with the environment holistically (cf. Caracciolo 2014: 19). This distinction is useful, as it provides the metaphorical and conceptual basis for the two approaches to ←123 | 124→literature we have dealt with so far: efferent reading, such as ‘rational’ narratological analysis, is based on schema theory and the computational model of cognition. Aesthetic reading, however, proposes a holistic approach that is more in line with embodied cognition and enactivism.

The next chapter introduces schema theory and its view of the brain as a ‘file clerk’ that opens folders retrieved from a vast library, based on incoming ‘requests’, uses the available information to make sense of input and updates the files to match the new situation. For this system to work, we need to have representations of everything that exists in our brains: “Whenever we try to deal with any aspect of the world in any way, we necessarily form a model of that aspect of the world” (Hogan 2003: 40). In Story Logic, David Herman relies on exactly the same mechanism to conceptualise

… narrative understanding as a process of building and updating mental models of the worlds that are told about in stories. In other words, story recipients, whether readers, viewers, or listeners, work to interpret narratives by reconstructing the mental representations that have in turn guided their production. (2002: 1)

Norman N. Holland calls this the “bi-active model of reading”, as the input determines which folders are opened, but the content of the folders determines our responses. Holland, who is a strong proponent of a reader-active model, calls this the “compromise position” (2009: 175), as the text is still assumed to dominate both perception and cognition. During the heyday of reader-response criticism he found Iser’s approach to reading too unidirectional and mechanistic (cf. Iser, Holland & Booth 1980: 58–9). While it is true that Iser stresses the overdetermination of literary texts, often to the extent of Meir Sternberg (1978) or David Bordwell’s (1985) narratological approaches, his model of meaning-making is more complex than that and mirrors Fauconnier and Turner’s conceptual integration theory, which is going to be a major argument towards the end of this part.

Schema theory developed out of joint research programmes on artificial intelligence and human cognition and led to the complex conceptual metaphor that the brain is a computer. As the example of bottom-up and top-down ‘processing’ demonstrates, we have become so accustomed to these metaphorical entailments that we take them to be literal – a phenomenon we are going to encounter again and again: we use the language of computers to refer to organic processes which have little to do with a binary code or hard drives. Our own interactions with computers are highly metaphorical, precisely because there is no natural similarity. We talk to them as if they were people, we arrange folders on our ‘desktop’, we throw old files into a bin, and we even believe that there ←124 | 125→are photos stored in neat boxes inside the hard drive. All of these things do not exist literally, of course, but we prefer the illusion that they do (cf. Fauconnier & Turner 2003: 22–24, 131).

Neither do computers have sense organs nor experiences, so programmers had to create schematic representations of objects, locations and procedures to make it possible to, first of all, identify them, and, secondly, interact with them. This approach assumes that our interactions with the environment are equally based on the conscious application of models and schemata. Learning and reading are thought to be based on the retrieval and storage of knowledge. New information is added to a vast semantic network where terms are organised in(to) hierarchies. The problem here is that this computational approach relies exclusively on semantic memory and tends to disregard three important resources that humans have: personal experiences, embodied cognition and emotions. Based on Endel Tulving’s research (1972) Groome makes

… a distinction between episodic memory, which is our memory for events and episodes in our own lives, and semantic memory, which is essentially our general knowledge store. Perhaps the most important difference between these two memory systems is that episodic memory involves the retrieval of a personal experience associated with a particular context (i.e. the place and time when it occurred), whereas semantic memory involves the retrieval of facts and information (such as the meanings of words), which are not attached to any particular context. (2014b: 177)

It is easy to see how this distinction relates to John Dewey’s emphasis on experience and Louise M. Rosenblatt’s discrimination between efferent and aesthetic reading and why narratologists, such as David Herman, are more in favour of schema theory. In the former case, reading is conceived of as a dynamic transaction with the text, based on a holistic approach, for which all human resources are needed: “Everything that we know and believe is organized into a personal theory of what the world is like, a theory that is the basis of all our perceptions and understanding of the world, the root of all learning, the source of hopes and fears, motives and expectancies, reasoning and creativity” (Smith 2004: 14). This equally applies to reading, which is more than an extraction of information: “Reading […] is best regarded as something done by people rather than by brains” (2004: 11).

Accordingly, chapter 3 addresses the importance of emotions and empathy, which are ignored in the computational model for obvious reasons. In Cognitive Poetics, Peter Stockwell begins chapter 11 with a surprising recontextualisation of his efforts up to this point:

←125 | 126→

The experience of literature, as described so far throughout this book, is one of rational decision-making and creative meaning construction. However, reading literature can also often be an emotional process, a felt experience, even offering a bodily frisson of excitement and pleasure, the prickling of the hairs on the back of your neck and a line or an idea or a phrase or an event that makes you catch your breath, and remember it for a long time afterwards. (2002: 151)

Yet, readers’ emotions, life experiences and entanglements with narratives provide more than passing sensations: they are constitutive of reading comprehension itself. Without some degree of empathy, characters would never come alive. Not least due to Suzanne Keen’s seminal book on Empathy and the Novel (2010) has literary studies rediscovered the emotions of readers. Following Keen, I present a systematic exploration of empathy in its various facets – eight, to be precise (cf. Batson 2009) – to shed some light on a concept that is not only central to literary reading, but also quite ambiguous.

The fourth chapter introduces embodied cognition, which proposes that humans intuitively understand new situations based on previous experiences in similar contexts. Faced with a hammer, we may remember specific situations that carry special autobiographical meaning. Otherwise, we have an immediate sense of its functional uses – called affordances – from handling different versions of this tool. In every single case, direct experience provides the basis for out interaction with the object and not abstract knowledge (cf. Evans & Green 2006: 241; Lakoff 1990: 50–2). This is how Antonio Damasio describes this phenomenon in The Feeling of What Happens:

we store in memory not just aspects of an object’s physical structure – the potential to reconstruct its form, or color, or sound, or typical motion, or smell, or what have you – but also aspects of our organism’s motor involvement in the process of apprehending such relevant aspects: our emotional reactions to an object; our broader physical and mental state at the time of apprehending the object. As a consequence, recall of an object and deployment of its image in mind is accompanied by the reconstruction of at least some of the images which represent those pertinent aspects. (2000: 183)

The same applies when we meet a person: very little conscious information comes to mind. At best, we remember the previous encounter or something we wanted to say or do. We are more likely to have an instant understanding of what is appropriate behaviour and how we are supposed to feel about this social encounter. At best, information always comes as a package that includes emotions and evaluations, especially where people are concerned. Thus, embodied cognition is a more recent iteration of Dewey’s basic idea that cognition and experience are inseparable. This equally applies to reading and our encounters with ←126 | 127→characters. Instead of creating ‘fact files’ that are filled with abstract data, we learn to read their entanglements, especially personal relationships.

Embodied cognition has diversified into three major strands that try to explain how we are able to understand other people: first, there are researchers who assume that we gradually develop our own folk psychology that helps us read other human beings. This is known as ‘theory theory’. It is the most ‘narratological’ of the three, as it tends to associate body codes with specific meanings. Secondly, many cognitive psychologists (cf. e.g. Oatley 2016, Green 2004) believe that we run a quick internal simulation – another metaphor – to compute how the other person feels and thinks. This is known as ‘simulation theory’. Both of them are frequently taken together and referred to as ‘Theory of Mind’ or “mind-reading” (Zunshine 2006: 7). A third group, the enactivists, who are philosophers rather than psychologists, attempt to trace this ability to repeated encounters with similar social situations and cultural training, which neither requires a theory – which is considered to be too close to a mental model approach – nor a simulation, since our brains are not computers into which we feed data and wait for the results. Despite these differences, they all share a strong concern with social interactions and the ability to ‘read’ other human beings. As a cognitive approach to reading literature, Theory of Mind foregrounds the importance of characters and privileges readers’ understanding of social minds (cf. Palmer 2004, 2010) over the traditional concerns of classical narratology.

In the last two chapters, the focus shifts to cognitive linguistics. The centrality of conceptual metaphors has already become apparent with the example that reading is a journey, but this is just the tip of the iceberg: to understand how comics encode experiences we also need to look at image schemas and conceptual metonymy. Finally, I want to demonstrate how Fauconnier and Turner’s theory of blending (2003) can be understood as a meaningful conceptualisation of reading, especially through a discussion of Barbara Dancygier’s The Language of Stories (2012), which represents an interesting point of comparison for Iser’s model. Since this has been a very swift introduction to a rekindled interest in the reader, it is now necessary to look at these theories in greater detail.

3.2 Mental Models

In their seminal study Scripts, Plans, Goals and Understanding: An Inquiry into Human Knowledge Structures (1977) Roger C. Schank and Robert P. Abelson offer a rationale how “a convergence of interests at the intersection of psychology and artificial intelligence” is motivated by similar concerns in the two fields: “What is the nature of knowledge and how is this knowledge used? These questions ←127 | 128→lie at the core of both psychology and artificial intelligence” (1977: 1). While the questions may have been strikingly similar, the observed phenomena were clearly not and led to curious generalisations: “For both people and machines, each in their own way, there is a serious problem in common of making sense out of what they hear, see, or are told about the world” (1977: 2). Retrospectively, this statement may seem odd, but it is a clear indication of how the two disciplines were inspired by each other: “If we understood how a human understands, then we might know how to make a computer understand, and vice versa” (1977: 8). Unfortunately, this influence tended to be one-sided at times:

we both believe that we need computers as the metaphor in terms of which we create our theories and as the arbiter of the plausibility of our theories. There is such a range of problems and procedures involved in the understanding process that to not use a computer is simply not to know whether what you are theorizing about could ever possibly work, let alone be right. (1977: iv)

This is exactly the same problem as looking at reading from the point of view of holistic understanding or advanced narratological analysis. What seems to be a straightforward process that most children can eventually master, becomes associated with “such a range of problems and procedures” (1977: iv) that the theory of acquisition for the first is increasingly bogged down by obstacles imported from a different discipline. Perception in general, and reading in particular, does not have to be extremely complicated just because it is difficult to describe in technical terms or to teach a machine how to do it. This is one of Frank Smith’s central arguments: “We live in an enormously complex and complicated world, but the times when individuals are actually confused, even babies, are remarkably few” (2004: 3; see also 23).

In his widely acclaimed Story Logic (2002), David Herman sets up the same trap for himself. He begins the book in the following manner: “Understanding long, detailed, and formally sophisticated literary narratives is for many people a natural, seemingly automatic process. Early on, however, artificial intelligence researchers showed that enormously complex linguistic and cognitive operations are required to generate or comprehend even the most minimal stories” (2002: 1). And then he proceeds to introduce a system of story comprehension that may be brilliant in its depth of analysis, but is as convoluted as any AI researcher could dream of. Herman starts from a very simple premise: we have to keep track of the development of characters as it is their motivations that drive the narrative. Yet, as a narratologist, he is more interested in “mapping the trajectories of individuals and objects as they move or are moved along narratively salient paths” (2002: 8), for which he has to establish the “principles for narrative ←128 | 129→microdesigns”, which “include coding strategies used to apportion particular facets of storyworlds into states, events, and actions” (2002: 6). These actions are then broken down into the “elements of the canonical description of an action” (2002: 62), of which he lists five major types, three of which have subcategories. Herman becomes so entangled in the details of his own classificatory system that he loses track of his initial aim to establish narrative as “a basic and general strategy for making sense of experience” (2002: 24).

Another aspect of this confusion of narratological analysis with reading is the idea that readers “reconstruct the storyworlds encoded in narratives” (Herman 2002: 5). Herman explains the process in the following way: “storyworlds are mental models of who did what to and with whom, when, where, why, and in what fashion in the world to which recipients relocate” (2002: 5). He believes that we ‘transport’ to the storyworld and then keep track of everything that is going on by updating our files – mental frames – with every new bit of information. This is incompatible with gestalt psychology and reader-response criticism, which both favour a holistic approach and readers’ intuitive meaning-making (cf. Miall 2006: 292). Based on their empirical research Gail McKoon and Roger Ratcliff developed a minimalist hypothesis in the 1990s according to which System 1 (cf. Kahneman 2012) operates locally as long as possible:

We claim that there is only minimal automatic processing of inferences during reading. Our hypothesis is that readers do not automatically construct inferences to fully represent the situation described by a text. In the absence of specific, goal-directed strategic processes [e.g. analytical tasks], inferences of only two kinds are constructed: those that establish locally coherent representations of the parts of a text that are processed concurrently and those that rely on information that is quickly and easily available. (1992: 440)

In other words, readers do not (re)construct storyworlds, the correct chronological order of events or character biographies automatically – unless they are foregrounded in the narrative itself and have direct relevance. Then we notice them and ‘mind’ in Dewey’s sense. As long as readers can make sense of a scene, there is no need to over-interpret, or, to be more precise, to interpret at all, if we understand interpretation as a conscious (System 2) operation. A coherent understanding of text does not require sophisticated analysis (cf. McKoon & Ratcliff 1992: 456; Smith 2004: 87, 96).

McKoon and Ratcliff offer a simple example: “Mary stirred her coffee” (1992: 457). The minimalist hypothesis suggests that we do not complete the picture by adding a spoon, which is the most important object in this scenario. As long as we have a general understanding that makes sense to us, there is no need to elaborate. A complete misreading of Iser would be to assume that he meant ←129 | 130→gaps in the situational model (cf. McKoon & Ratcliff 1992: 458). The generative power of gaps has to do with readers’ understanding of how the foregrounded elements of the narrative can be put into meaningful relations to each other. There is a significant difference between acknowledging the usefulness of generic schemata to make sense of basic situations, which corresponds to Iser’s automatic consistency-building, and claiming that we have to recreate all elements of the storyworld as a mental model in our brains. Tests have shown that readers need a lot of cognitive effort to trace the placement of objects in the virtual space of the story, reconstruct the floorplan of houses based on a reading, or any similar task that exceeds an understanding of what has been explicitly foregrounded (cf. McKoon & Ratcliff 1992: 461; see also Johnson-Laird 1990: 158–62; Emmott 2004: 46–50). Criticising Philip Johnson-Laird for his “equation of natural story-reading with ‘a superficial understanding’ ”, Catherine Emmott points out that “salience and reading purpose” have to play a much larger role in cognitive approaches to reading: “The majority of readers may not be concerned about the precise positions of objects in a room because this is generally not the main point of reading a passage like this” (2004: 47). What is required is an understanding of “the actions of the central characters and an overall impression of the location” (2004: 47), unless, of course, the narrative specifically invites the readers to pay special attention to details.

In her article “Cognitive Maps and the Construction of Narrative Space” Marie-Laure Ryan perfectly states the case:

It takes a specific agenda – such as the present project – to attempt the systematic reconstruction of the ‘textually correct’ map of a fictional world. It was only on my third reading of Chronicle of a Death Foretold that I reached what I hope is a reasonably complete and accurate representation of the topography of the novel. My first reading was a reading for pure pleasure. (2003: 217–8)

As McKoon and Ratcliff argue, the point to be made is not that it cannot be done through goal-directed strategic processes, as Ryan illustrates, but that highly abstract thinking and world-building are not a natural part of reading. Like Ryan, Emmott makes a difference between herself as “a first-time reader of the story” and as “an analyst” (2004: 256). She also offers an interesting observation on the inclusion of maps as peritexts of prose fiction: “Occasionally novels do include lay-out drawings and maps, but presumably readers consult these diagrams when necessary rather than memorize them. The fact that these diagrams are deemed necessary in such cases perhaps indicates that the narrative text is not a good means of conveying this information” (2004: 49). In her article “ ‘Situated Events’ in Fictional Worlds: The Reader’s Role in Context Construction” Emmott ←130 | 131→offers an extensive argument why the characters’ fictional environments cannot be separated from their experiences:

The idea of the reader mentally controlling ‘contextual frames’ is rather different from traditional notions of ‘place’ in narratology (e.g., Lodge, Rimmon-Kenan). ‘Place’ is usually thought of as the location itself and is conveyed by descriptions (e.g., Chatman, Hamon). In the model described in the previous section, information about a particular place would be stored in a location representation, whereas a contextual frame provides information about the ‘placing’ of characters and objects in a spatio-temporal configuration. The notion of contextual frames provides a model of the reader actively tracking the dynamics of the fictional world and assembling a context from and around the events which occur. Also, rather than drawing a dividing line between the characters and the location, this puts the emphasis on each character being surrounded by the people and the location. This provides a view of our physical environment which is social in nature – our actions can only be fully understood by taking into account who is around us.” (1998: 191)

Manfred Pfister uses the term “configuration” to refer to “the dramatis personae that is present on stage at any particular point in the course of the play” (2000: 171), but Emmott extends this concept to all salient elements and their interrelations. An interesting parallel between Iser, Emmott and Pfister is how they conceptualise meaning-making as drawing different contextual frames or salient elements together for mutual illumination: “The identity of a dramatic figure takes shape and evolves in the series of configurations in which it participates, and the contrasts and correspondences that develop between one particular figure and the others become clear when they are meaningfully juxtaposed on stage” (Pfister 2000: 172).

Like Herman, teachers may be tempted to ask “who did what to and with whom, when, where, why, and in what fashion in the world” of the narrative (2002: 5), but literature deals in entanglements and readers do not primarily store factual information unless foregrounded. In a network model of cognition “the connections between items of information are regarded as being as important as the information itself” (Emmott 2004: 51), a point that remains underspecified in Herman’s Story Logic, despite claims to the contrary:

Interpreters of narrative do not merely reconstruct a sequence of events and a set of existents but imaginatively (emotionally, viscerally) inhabit a world in which, besides happening and existing, things matter, agitate, exalt, repulse, provide grounds for laughter and grief, and so on – both for narrative participants and for interpreters of the story. More than reconstructed timelines and inventories of existents, storyworlds are mentally and emotionally projected environments in which interpreters are called upon to live out complex blends of cognitive and imaginative response, encompassing ←131 | 132→sympathy, the drawing of causal inferences, identification, evaluation, suspense, and so on. (2002: 16–17)

To put this whole argument in a nutshell: while our reading is heavily schema-based when System 1 engages in automated consistency-building and finds instant connections on a localised level, we rely far less on global inferences, cognitive storyworld models or more abstract schemata – like narrative or genre models – to make sense of specific scenes. Here, contextual frames play a far more important role, which are mental representations of scenes and character relationships rather than facts.

Yet, even with the most mundane, everyday situations, the notion that our thinking is based on abstract, generalised mental models is counterintuitive at first, as all of our experiences result from specific contexts. Accordingly, when Schank and Abelson developed their theory of schemata, they had to start with episodic memory:

The form of memory organization upon which our arguments are based is the notion of episodic memory. An episodic view of memory claims that memory is organized around personal experiences or episodes rather than around abstract semantic categories. If memory is organized around personal experiences then one of the principal components of memory must be a procedure for recognizing repeated or similar sequences. (1977: 17–18)

In other words: to avoid constant cognitive overload and to allow for quick reactions during life-threatening circumstances, humans had to develop a system that relied on the recognition of and automated response to general patterns (System 1) and reserved conscious processing (System 2) for deviations from the norm.

Some episodes are reminiscent of others. As an economy measure in the storage of episodes, when enough of them are alike they are remembered in terms of a standardized generalized episode which we will call a script. Thus, rather than list the details of what happened in a restaurant for each visit to a restaurant, memory simply lists a pointer (link) to what we call the restaurant script and stores the items in this particular episode that were significantly different from the standard script as the only items specifically in the description of that episode. (1977: 19; see also 37)

Since these ‘generalized episodes’ are cultural models that are shared by a community (cf. Stockwell 2002: 33), it is possible to start with the pointer and then only narrate what was noteworthy about a situation. Many conversations start with: “You cannot imagine what happened to me at/during …”. Thus, the schema is evoked first and then we just report the unusual circumstances, which requires some skill in making the details cohere without repeating large parts ←132 | 133→of the routine (cf. Schank & Abelson 1977: 45). This applies to oral storytelling as much as to other types of communication: “People, in speaking and writing, consistently leave out information that they feel can easily be inferred by the listener or reader” (1977: 22; see also 23). This is different with young children who still have to learn how to abstract and tend to narrate entire episodes: “A child must learn that experiences that differ by a few small items are in fact best handled by one script. Early on in the script acquisition process, children do not realize this, and often see no similarity in events that seem nearly identical in form to an adult” (1977: 232).

Most of the gap-filling that daily life requires of us is pretty mundane. Therefore, art focuses either on the unusual or a fresh look at familiar things that deserve closer attention. In this sense, reading relies to a large extent on schemata for Iser’s automated process of consistency-building, but the point of doing that is conscious attention to that which is artistically foregrounded and special. Accordingly, schemata are relevant standardised models against which the literary comes to the fore. This is also the central difference between teaching language and literature: while formulaic, schema-based teaching makes sense to enable students to master standard situations and text types, an engagement with literature requires more than identifying genre markers or narratological categories.

Schank and Abelson’s theory also clarifies that we approach situations holistically, that we always start with general orientation and overall meaning and then attend to the particulars: “In understanding it seems doubtful that people first do a syntactic analysis without recourse to meaning and than [sic] look at the meaning. People understand as they go” (1977: 16). In daily routines, we basically assume that everything is going according to plan and that the new situation is going to be as comfortably boring as the few hundred times before (cf. 1977: 67). In this sense, we use schemata to project what is going to happen. In an educational setting, predictions are a natural and important part of reading and can help, as in real life, to establish some basic orientation. When teachers come back to the initial list of predictions after a reading sequence, students should be praised for noticing deviations from the generic patterns and not for getting everything right in the first place. This is another reason why rereading and the analysis of specific scenes is so important, to get a sense of how the work of art is more than its generic backbone.

The concept of schemas predates Schank and Abelson’s approach by a few decades, of course, and is usually traced back to Frederic Bartlett’s Remembering (1932) and Jean Piaget’s observations on children’s developmental stages in The Language and Thought of the Child (1926). This is how Piaget explains the ←133 | 134→holistic approach to cognition in the context of using tachistoscopes in experimental research, which are devices that showed test subjects images or words for milliseconds:

Recent research on the nature of perception, particularly in connexion with tachisto[s];copic reading, and with the perception of forms, has led to the view that objects are recognized and perceived by us, not because we have analysed them and seen them in detail, but because of “general forms” which are as much constructed by ourselves as given by the elements of the perceived object, and which may be called the schema or the gestaltqualität of these objects. (1926: 131)

Bartlett was equally influenced by Gestalt psychology, as was Iser many decades later. The former’s schema theory is described by Groome in the following manner:

we perceive and encode information into our memories in terms of our past experience. Schemas are the mental representations that we have built up from all that we have experienced in the past, and according to Bartlett we compare our new perceptual input with our schemas in an effort to find something meaningful and familiar. Any input which does not match up with existing schemas will either be distorted to make it match the schemas, or else it will not be retained at all. (2014b: 161–2; see also 2014a: 8)

Bartlett never tires of stressing that these schemas are not fixed forms, but constantly evolving patterns or sets of various experiences that are connected by a common interest. As much as schemas shape our perception, they are equally modified by new experiences. Bartlett discusses the conceptual metaphor the brain is a storehouse to highlight its shortcomings:

In any case, a storehouse is a place where things are put in the hope that they may be found again when they are wanted exactly as they were when first stored away. The schemata are, we are told, living, constantly developing, affected by every bit of incoming sensational experience of a given kind. The storehouse notion is as far removed from this as it well could be. (1964: 200)

In essence, we learn about the world by updating these models. Here is Piaget’s version of the same idea, which he calls ‘assimilation’:

… reality data are treated or modified in such a way as to become incorporated into the structure of the subject. In other words, every newly established connection is integrated into an existing schematism. According to this view, the organizing activity of the subject must be considered just as important as the connections inherent in the external stimuli, for the subject becomes aware of these connections only to the degree that he can assimilate them by means of his existing structures. (Piaget & Inhelder 2000: 5)

←134 | 135→

In contrast to ‘assimilation’, the process of ‘accommodation’ involves “the modification of internal schemes to fit reality” (2000: 6), which leads to a significant re-structuring of existing patterns.

Since “all mental organization is schematic in nature” (Mandler 1984: 2), according to Jean Matter Mandler, there have been attempts to classify mental models. In Cognitive Science, Literature, and the Arts (2003) Patrick Colm Hogan uses a distinction between “representational schemas”, or schemas in a narrower sense, which he associates with the mental lexicon, taxonomies and lists of features for each entry, and “procedural schemas” (2003: 44), which are usually called scripts or “event schemas” (Mandler 1984: 13), which Mandler identifies as “the stereotypical knowledge structures that people have acquired about common routines” (1984: 75). Here is Schank and Abelson’s definition:

A script is a structure that describes appropriate sequences of events in a particular context. A script is made up of slots and requirements about what can fill those slots. The structure is an interconnected whole, and what is in one slot affects what can be in another. Scripts handle stylized everyday situations. They are not subject to much change, nor do they provide the apparatus for handling totally novel situations. Thus, a script is a predetermined, stereotyped sequence of actions that defines a well-known situation. Scripts allow for new references to objects within them just as if these objects had been previously mentioned; objects within a script may take ‘the’ without explicit introduction because the script itself has already implicitly introduced them. (1977: 41; see also Stockwell 2002: 77; Smith 2004: 21–22)

Schank and Abelson’s most famous example is the restaurant script (cf. 1977: 42–6), which they reference throughout the book: “When we refer to ‘the’ restaurant script, therefore, we are relying on those stereotyped details which are culturally consensual” (1977: 55). The most important aspect of our reliance on scripts is that they belong to the automated processes of System 1 and save a lot of energy by not requiring conscious attention: “The waitress typically does what the customer expects, and the customer typically does what the waitress expects. There is great social economy when both parties know the script because neither party need invest effort deciding what the actions of the other mean and how appropriately to respond” (1977: 61). When our “predictive powers” (1977: 45) fail, however, it is fascinating to see how easily people get confused by minor adjustments to the familiar patterns, be it driving on the left side or waiting to be seated in restaurants. Since our understanding of literature is based on real-life experiences and the other way round, there is a feedback loop through which art and life inform each other: “Scripts are important not only in guiding our own action, but in understanding other people’s actions and reports of actions, including those reports that appear in literature” (Hogan 2003: 45).

←135 | 136→

While Vladimir Propp’s Morphology of the Folktale, which was published in the Russian original in 1928, attempted to schematise the text-immanent structures of fairy tales, the 1970s and 80s witnessed a veritable boom of explaining the comprehension of narratives in terms of mental models. Mandler’s Stories, Scripts, and Scenes: Aspects of Schema Theory (1984) is a typical example which proposes an exact analogy between acquiring scripts through (and for) social encounters and developing story schemas by being exposed to narrative texts. Early on, she makes an important distinction between text-immanent and cognitive structures:

A story grammar is a rule system devised for the purpose of describing the regularities found in one kind of text. The rules describe the units of which stories are composed, that is, their constituent structure, and the ordering of the units, that is, the sequences in which the constituents appear. A story schema, on the other hand, is a mental structure consisting of sets of expectations about the way in which stories proceed. The close connection between a story grammar and a story schema arises from the fact that the story schema is a mental reflection of the regularities that the processor has discovered (or constructed) through interacting with stories. (1984: 18).

In contrast to a feature-based narratological analysis, Mandler states that “much of the work of cognitive structures goes on beyond the reach of consciousness” (1984: xi), which returns us to Iser’s idea of automated consistency-building and Kahneman’s System 1 (cf. Mandler 1984: 32–5, 108; Kahneman 2012). According to this logic, “[w];hat is consciously noticed is a discrepancy from the normal values, the violation of an expectation” (1984: 35; see also 101; Stockwell 2002: 20), which brings Mandler’s approach again in line with Iser and the concepts of selection, overdetermination and defamiliarisation: “The schema prepares the person to see certain kinds of things; consequently, little attention need be paid to those things that match the expectations, leaving attentional resources free to devote to the more unusual, and therefore more informative, items” (1984: 105; see also 26, 103).

In this sense, genre competence, which is a related concept, provides an essential first orientation and a meaningful way into a story, until the more narrative-specific elements gain prominence and supersede the generic schema. Retrospectively, a genre label may help to find back into a story and remember the details and deviations by working from a common ground towards the more unusual aspects. However, “with the passage of time, recall becomes more dependent upon a generic knowledge representation than on the specifics of individual statements” (Mandler 1984: 73). This is a point that Mandler proved empirically: “From these various recall studies, we have found support for the presence of an idealized form for a particular kind of story. The data provide ←136 | 137→evidence both for the constituents of a story schema and for the sequences in which those units are strung together” (1984: 50; see also 105). Her research raises the question how the specific narrative relates to the generic model, the ‘idealized form’, which can be answered by recourse to prototype theory and mental classifications.

Both are essential to schematisation in the context of the mental lexicon, as words and concepts are not stored separately, but in an organised manner: “we can say that prototypicality is the basis of categorisation, with central examples acting as cognitive reference points in the middle of a radial structure” (Stockwell 2002: 29). With scripts, we have seen that objects and roles are reduced to their generic functionality and have a fixed place in the overall pattern. Even in completely new settings it is easy to identify who is who and what is what without resorting to conscious analysis. This is possible because we abstract a prototypical situation from countless specific experiences that is universal enough to guide us through as many individual contexts as possible and allows us to attend to the specific elements that are new. “Prototypes are, basically, standard cases” (Hogan 2003: 45) and it takes children a long time to learn these cultural classifications by testing labels and finding out the optimal degree of specificity and/or generality. Stockwell explains these ‘basic level’ prototypes in the following manner:

The basic level tends to be the level at which we most commonly interact on a human scale with the category. We distinguish basic level objects at the point where they seem to have the most discontinuities with other objects in the world. Terriers are not as different from collies as dogs are different from cats. The basic level is also where most of the attributes of a category are optimally available – we tend to have more of a sense of ‘dogginess’ than ‘collie-ness’ or ‘mammal-ness’. These hierarchies of superordinacy and subordinacy are what allow us to use and recognise over- and under-specificity. […] Recognising categories seems to be a two-stage process, involving a holistic perception of the category as an object (a ‘gestalt’ whole) followed, if necessary, by an analytical decomposition of the object into separate chained subtypes or attributes. (2002: 31)

There are three important things to notice in this context: prototypes provide an important, almost instantaneous first orientation that is sufficiently accurate. In most contexts an ‘analytical decomposition’ is neither possible nor necessary, as the main aim of operating with prototypes is automation. Scientists tend to break down – ‘analyse’ – objects into atomistic features or parts, but ‘reading’ in the broadest possible sense is about holistic perception, as Stockwell argues. Prototypes are necessarily stereotypes and there is no escape from them. All our thinking is based on preconceived or prejudiced ideas:

Our knowledge about an object or classes of objects, about an event or classes of events, about personality traits and social norms, can all be considered as small networks of ←137 | 138→information that become activated as we experience these things and that function according to certain schematic principles. (Mandler 1984: 2–3)

Therefore, Mandler describes schemas “as sets of expectations” (1984: 13) as we do not objectively register what is out there, but rather judge to what extent new information conforms to our preconceived notions. Smith argues that predictions are central to our lives as human cognition is geared towards projection and extrapolation:

Everyone predicts – including children – all the time. Our lives would be impossible, we would be reluctant even to leave our beds in the morning, if we had no expectation about what the day might bring. We would never go through a door if we had no idea of what might be on the other side. And all our expectations, our predictions, can be derived from only one source, our theory of the world. We are generally unaware of our constant state of anticipation for the simple reason once again that our theory of the world works so well. (2004: 23; see also 25; Gombrich 2014: 155, 170–1, 254)

Still, there is hope to enrich engrained ideas with new and more diversified perspectives or emotions, so that System 1 may eventually call forth different associations: “The junction of the new and old is not a mere composition of forces, but is a re-creation in which the present impulsion gets form and solidity while the old, the ‘stored,’ material is literally revived, given new life and soul through having to meet a new situation” (Dewey 2005: 63). Schemas have to be prejudices to work effectively, but they can be reshaped over time in light of new experiences.

A second and related point concerns the cultural context: “The prototypical man for any given person will involve average properties, not of all men, but of men who are highly salient in that person’s experience” (Hogan 2003: 46), such as ideas about height, skin colour, overall build, clothes, posture etc. Prototypicality is culture-specific and thus malleable over time and especially in those areas in which we are still willing to learn. Travelling, living abroad or emigrating may pose a challenge to our established schemas, so that successful integration has to involve the modification of numerous prototypes. This does not primarily apply to details, but to basic forms of interaction.

A third and final point concerns the differentiation between prototype, example/exemplum and exemplar (cf. Hogan 2003: 46–7; see also Evans & Green 2006: 249). A prototype, as we have seen, best captures the sense of ‘dogginess’ or ‘restaurantness’ without being a specific example. Thus, it represents an abstract centre to which examples are related at various distances according to family resemblances (cf. Lakoff 1990: 12–16): “It seems that our cognitive system for categorisation is not like an ‘in or out’ filing cabinet, but an arrangement of elements ←138 | 139→in a radial structure or network with central good examples, secondary poorer examples, and peripheral examples. The boundaries of the category are fuzzy rather than fixed” (Stockwell 2002: 29). This categorical system is ever-changing, especially when looking at exemplars, which are supposed to be the ‘central good examples’ of a specific type and may set new standards. Summarising recent research, Richard Gerrig suggests that we do not store a single prototype, but rather work with several exemplars as points of reference (cf. 2011: 40). This may be more appropriate in the context of literature, as a focus on the most prototypical examples as role models would be problematic. ‘Genre fiction’ is a derogatory term that is used for formulaic narratives that excessively follow a fixed pattern. When works of art emulate an exemplar too closely, they become clones or copies and may be seen as cases of plagiarism rather than works of creativity. In Hogan’s jazz example creativity is associated with the playful de- and re-construction of generic models (cf. Hogan 2003: 70–86). Artists have to be both skilful and familiar with the conventions to be able to play with them in a seemingly effortless manner. Readers acquire genre awareness by comparing and contrasting different texts, as exemplars are – by definition – not typical (cf. Stockwell 2002: 30). Therefore, the choice of text ensembles (cf. Delanoy 2015: 20, 24, 29) for the classroom plays an equally important role in genre studies as it does in inter/transcultural learning.

Summarising the argument so far, there are two basic types of memory: semantic and episodic memory. The first is based on generic knowledge structures that play a crucial role in automated responses (System 1) and can be further subdivided into representational and procedural schemas. The second type encompasses specific personal memories. In cognitive poetics there is a strong bias in favour of the first, in aesthetic reading of the second. When students are encouraged to establish a personal connection with a text, this is much more likely to be based on idiosyncratic experiences and not on culturally transmitted generic conventions, such as restaurant scripts. What episodic memories share with scripts is their complex integration of spaces, objects, roles and procedures. However, while scripts remain at the generic level, they become a background in episodic memories against which the ‘tellable’ parts emerge. All these terms and classifications are metaphoric in nature, as it is still not entirely clear how memories are stored and retrieved. A practical solution that is also intended to bridge the gulf between cognitive and experiential approaches is the introduction of what Lakoff calls “idealized cognitive models, or ICMs” (1990: 68) and which Stockwell understands as “structures with which we organise our knowledge. Cognitive models consist of relations between categories, set up socially, culturally, and on the basis of individual experience, as ←139 | 140→our means of understanding and negotiating the world and our lives through it” (2002: 33). Here we have an important step forward in that humans are no longer said to operate with particulars, but meaningful connections between semantic entries. Before we return to mental spaces and their networked interrelations in chapter 5, we look at two phenomena in which cognition is conceptualised as a holistic experience: empathy and enactivism. They relate to two central questions that every teacher of literature has to face: to what extent can students identify with the social world and especially the characters of narrative fiction? And, secondly, do we learn to read and manage social encounters through a bottom-up process of building competence via practice and intuition or through a top-down process of situational analysis and applying schemata?

3.3 Emotions & Empathy

3.3.1 The Feeling of What Happens

While the previous chapter highlighted the representation and classification of generic knowledge, this one shifts attention to the interrelatedness of body and mind, or embodied cognition, which proposes a much closer tie between cognition, emotions and evaluations. Antonio Damasio’s The Feeling of What Happens: Body, Emotion and the Making of Consciousness may serve as a useful point of departure for an exploration of the key concepts of cognitive studies after the dominant paradigm of the computational model began to fall into disrepute. Damasio deplores the fact that in first-generation cognitive science “the brain remained consistently separated from the body rather than being seen as part of a complex living organism. The notion of an integrated organism – the idea of an ensemble made up of a body proper and a nervous system – […] had little impact in shaping the standard conceptions of mind and brain” (2000: 40). To counter this misconception, he redefines the conscious mind as an exploratory tool that serves human beings in their interactions with the environment. “The brain is a creative system. Rather than mirroring the environment around it, as an engineered information-processing device would, each brain constructs maps of that environment using its own parameters and internal design, and thus creates a world unique to the class of brains comparably designed” (2000: 322). Instead of engaging in computational processes, such as assessments, calculations and simulations – which all conceptualise the mind as being turned inwards – Damasio conceives of the mind as being directed outwards (cf. 2000: 28–9). It is intimately involved in human experiences and interactions: “rather than concentrating resources on our inner states, it is perhaps more advantageous ←140 | 141→to concentrate one’s resources on the images that describe problems out in the world or on the premises of those problems or on the options for their solution and their possible outcomes” (2000: 29). In this sense, curiosity is built into the system and humans are explorers by design.

If teachers are willing to accept Damasio’s point of view, they are faced with three consequences that affect all aspects of their professional lives. First, thinking is not predominantly an independent manipulation of symbolic representations that are stored in the brain, but a key resource in a human being’s holistic engagement with an environment. Secondly, emotional responses and evaluations are directly tied to cognition (cf. 2000: 16). Thirdly, learning in Piaget’s sense of accommodation takes place by consciously noticing and directly interacting with specific objects. This may help to explain Damasio’s seemingly cryptic explanation of where the title of the book comes from: “the presence of you is the feeling of what happens when your being is modified by the acts of apprehending something” (2000: 10). Meaningful encounters leave a trace in episodic memory and thus change who we are: “Extended consciousness occurs when working memory holds in place, simultaneously, both a particular object and the autobiographical self, in other words, when both a particular object and the objects in one’s autobiography simultaneously generate core consciousness” (2000: 222). All memories are relational in this sense, as they come into existence through conscious interaction (cf. 2000: 20). Although Damasio acknowledges the role of schematic memory and automated processing, he shows more interest in subjective selves, autobiographical memory and an actively engaged mind. This can be explained based on his background in neurophysiological research, dealing with patients who suffer from an impaired consciousness (cf. 2000: 6) and whose ability to interact presents an important indication whether their minds can still reach out (extended consciousness) or are limited to core consciousness. For an empirical researcher such a situation requires a difficult triangulation between several factors:

Based on what we know about private human minds and on what we know and can observe of human behavior, it is possible to establish a three-way link among: (1) certain external manifestations, e.g. wakefulness, background emotions, attention, specific behaviors; (2) the corresponding internal manifestations of the human being having those behaviors as reported by that human being; and (3) the internal manifestations that we, as observers, can verify in ourselves when we are in circumstances equivalent to those of the observed individual. This three-way linkage authorizes us to make reasonable inferences about human private states based on external behavior. (2000: 83)

Despite these obvious limitations we keep “theorizing constantly about the state of mind of others from observations of behaviors, reports of mental states, and ←141 | 142→counterchecking of their correspondences”, which Damasio calls a “natural human ability” (2000: 83). Through years of training, we become fairly good at it. Since actions, reactions, emotions and thoughts are part of an integrated system, we learn to complete a fuller picture based on a few hints: “Just as the music you hear is the result of many groups of instruments playing together in time, the behavior of an organism is the result of several biological systems performing concurrently” (2000: 87). In short, even without a theoretical background in music or behavioural sciences, we can still identify a sour note or register that something is wrong.

According to this logic, mental states become embodied and we learn to ‘read’ what another human being is doing, perceiving, thinking and feeling, based on physical expressions: “Consciousness and mind […] are closely tied to external behaviors that can be observed by third persons” (2000: 12). Visual narrative media that may grant little access to characters’ thoughts and feelings rely on such an ability to ‘read minds’ based on social context and physical outward expressions, as we shall see in part 4. Like thoughts, emotional responses have to be seen as relational. They are not internal states or occurrences, but targeted at external objects and thus determine our interactions with them: “Emotion is critical for the appropriate direction of attention since it provides an automated signal about the organism’s past experience with given objects and thus provides a basis for assigning or withholding attention relative to a given object” (2000: 273). This (re)integration of thinking into a human being’s perception and experience is nothing new, as we saw with Dewey’s approach. This is how Bruner addresses a similar point in Actual Minds, Possible Worlds:

David Krech used to urge that people ‘perfink’ – perceive, feel, and think at once. They also act within the constraints of what they ‘perfink.’ We can abstract each of these functions from the unified whole, but if we do so too rigidly we lose sight of the fact that it is one of the functions of a culture to keep them related and together in those images, stories, and the like by which our experience is given coherence and cultural relevance. (1986: 69)

This also relates to the theory of social minds, that there is an “incontrovertible correlation between the private and the public” (Damasio 2000: 13). A lot of thinking takes place in a networked fashion, relying on other people and material anchors (cf. Hutchins 2005; Oatley 2013: 452).

For Damasio, emotions do not just occur at the same time as cognitive interaction takes place and thus may influence the outcome of the engagement, but they are directly tied and “integral to the processes of reasoning and decision making” (2000: 41; see also 43, 58). We do not extract factual information from ←142 | 143→experiences, but store them as bundles that also include our emotions, feelings and especially appraisals. In simple terms, emotions are the raw materials (core affects) on which our feelings are based. When we become aware of emotions, we attempt to rationalise, explain and label them. In case we are successful, we call these constructs our feelings. Appraisals are judgements of value (evaluations) in a particular context that are guided by emotions, but do not have to be conscious (cf. Oatley & Johnson-Laird 2014: 134–7). For example, we may instantaneously like people and have no idea why. When we respond positively to a piece of music (core affect), often subconsciously at first, e.g. by listening to it on the radio for some time without even noticing it, we later interpret this mood as enjoyment. Thus, it becomes a conscious appraisal of the music and stored in our memory alongside the title of the piece and/or the name of the artist(s). When someone finally decides to buy the CD, emotions have played a central role all along. Not surprisingly, most advertising tries to capitalise on this interconnection, often through a manipulation of potential customers’ passions, their nostalgia, sexual desire, pride, anger or a longing for social recognition.

The unavoidable omnipresence of emotions also returns us to Dewey’s concept of experience and the need to acknowledge students’ personal responses to texts in aesthetic reading. It may be prudent for professional narratologists to keep their feelings from interfering with their work, but in an educational setting, in which students are supposed to react to narrative texts – to enjoy literature, to become curious about characters and sympathise with them – it would be both counterintuitive and counterproductive to exclusively appeal to their rational minds. Emotions, as The Feeling of What Happens demonstrates, are a sign that people care and are actively – meaning cognitively – involved. Teachers may find themselves frustrated by first impressions that are coloured by strong emotions, but this is a much more appealing starting point than apathy and can lead to a productive return to the text.

Damasio differentiates between emotions, which are universal and accessible through physical responses to situations on public display (cf. 2000: 59, 73), and feelings, which he defines as private experiences and interpretations of such emotions (cf. 2000: 42). He observes that, etymologically, ‘e-motion’ refers to a movement, to “externalized behavior” (2000: 70), which we learn to read. Like factual information, we can recall, rationalise and newly appraise feelings, but also use them as the basis for future encounters with the same person or situation: “Well-targeted and well-deployed emotion seems to be a support system without which the edifice of reason cannot operate properly” (2000: 42). If we were not able to remember the specific ‘feel’ of social encounters, we would either not know how to behave the next time or cause irritation by completely ignoring ←143 | 144→‘what happened’. This necessity to read and project emotional states in everyday interactions, sometimes based on little evidence, lets us anthropomorphise animals or even inanimate objects. This explains why we can sympathise with things or machines as long as their outward expressions broadly resemble physical signs of emotions. The same logic applies to cartooning, which heavily relies on our ability to read intentions and emotions into highly abstract representations of characters.

Damasio argues that emotions are part and parcel of System 1 operations, which explains why we can make decisions before we become conscious of them. This may also mean that we feel/know what we have decided, but cannot rationalise and verbalise it yet: “Language – that is, words and sentences – is a translation of something else, a conversion from nonlinguistic images which stand for entities, events, relationships, and inferences” (2000: 107). Based on his research, Damasio argues that “there must be a nonverbal self and a nonverbal knowing” (2000: 108), which runs counter to the widespread belief that all cognition is based on language and rational thought. In a short chapter entitled “The Naturalness of Wordless Storytelling” (2000: 188–9) he observes that “[m];ovies are the closest external representation of the prevailing storytelling that goes on in our minds” (2000: 188). He postulates a correlation between how the viewer automatically integrates fragmentary ‘shots’ into a continuous action and how “the brain naturally weaves wordless stories about what happens to an organism immersed in an environment” (2000: 189). In both cases consistency-building is achieved without much conscious processing.

3.3.2 Types of Reading-Related Feelings

Having established that emotions and, in turn, feelings are closely tied to cognition, it is now time to look more specifically at the role of emotions during and after reading and their relevance for the literary classroom. There is a tendency to reduce emotional responses to literature either to empathy or to sympathy and then limit that again to a very specific type of behaviour, such as taking a character’s perspective or developing a pro-social attitude. Following Damasio’s lead, I would like to demonstrate that a full range of reader emotions is ever present.

The first important realisation is that empathy is just one of four categories of readerly feelings that David S. Miall and Don Kuiken identify in their essay “A Feeling for Fiction” (cf. 2002: 223; see also Kuiken et al. 2004: 174–5). These are obviously related and influence each other to varying degrees; yet, they differ enough to warrant separate introductions. The first group consists of evaluative ←144 | 145→feelings, which reflect readers’ attitudes towards a specific text, a genre or reading in general, based on the overall experience, such as satisfaction, enjoyment, a sense of accomplishment or even eagerness to read more by the same author. “Evaluative feelings emerge early within the reading event, with gradual adjustment throughout […], but they may affect readers’ moods – and their readiness to reread the text – for some time afterward” (Kuiken et al. 2004: 174). Thus, feelings of this type are blends or meta-feelings that compress individual experiences into a single overall impression. This already starts with the first emotional reactions to a text, which significantly influence readers’ attitudes and motivations throughout. The primacy effect is based on a blending phenomenon, as we begin to form initial concepts that determine our further reading experience. Whether that pattern is a general frustration with the difficulty of the text or intense excitement about the artistic rendering of a narrative, readers are likely to find more confirmation of their first impressions in the text. This is why students should get a chance to voice their concerns and exchange their views early on in the reading process.

Evaluative feelings are blends, or feelings about feelings, which means that they compress particular reactions into broader categories. In everyday conversations people are prone to share such global impressions, often exaggerating the responses through strong evaluative language. Thus, teachers are more than likely to encounter such responses at first. Quite a few students tend to love or hate a text, and find ‘everything’ inspiring, heart-warming, ‘totally’ boring or ‘completely’ exaggerated. Evaluative feelings are also very tempting for EFL students, as they only require a one-word answer. They sound like fundamentalist statements that are very far removed from the intricacies of the text that teachers would like to address rather sooner than later. Accordingly, evaluative feelings are often deemed inappropriate, understood as indications of the students’ lack of sophistication or appreciation and, therefore, ignored. However, since feelings are directly tied to students’ experiences of texts, they have to play a central role in classroom discussions. This is why I suggested using online discussion forums in the previous part. By collecting first responses that are more elaborate than single words and already contain a short explanation plus responses by other students, it is easier for teachers to prepare for the first meeting in class and channel the responses into productive discussions. Evaluative feelings are blends of the next three types of responses, which means that they can be traced back to their roots. As a consequence, students should be able – and actively encouraged – to decompress their generalisations, e.g. ‘boring’, and call forth more specific observations, e.g. that ‘the characters resemble stereotypes’, or even single incidences, such as that ‘the protagonist behaves in the most predictable manner ←145 | 146→in one particular scene’. This draws attention back to the text and makes it possible to (eventually) argue over some of these more specific points in pair, group or lockstep discussions.

Type 2 covers narrative feelings, which are emotional responses during the reading process that have an interpretative or evaluative function regarding situations (mood) and/or characters (empathy/sympathy). Since the importance of empathy vs. sympathy dominates discussions about emotional responses to literature, most of the latter part of this chapter is dedicated to a discussion of these phenomena. Richard Gerrig proposes a category of reactions that he calls “participatory responses”, which “covers all noninferential responses in the performance of narratives” (1998: 27) and places them halfway between this and the next type. Whenever readers respond physically (being at the edge of their seats; hiding under a blanket; biting their nails) and/or verbally (Stay away! Don’t go in!) to a scene, they are emotionally invested in the fate of the characters, but such reactions are also often deliberately set up and triggered by the narrative itself: “readers often experience suspense with respect to potential outcomes to which the characters are oblivious” (1998: 169). As a form of dramatic irony, this is such a widespread phenomenon that it has its own literary term. These ‘pre-programmed’, physical responses lead us to the next category.

The third group, called ‘aesthetic feelings’ by Miall and Kuiken, comprises fascination, interest, intrigue, surprise, shock, suspense, anticipation etc. These are evoked by “foregrounded structures” (2002: 224) or defamiliarisation and challenge the readers’ understanding of the text (cf. Oatley 1994: 58–9). They are based on stylistic devices or artful plotting in Meir Sternberg’s sense (cf. 1978), which entails the strategic suppression, manipulation and dissemination of information. For visual narratives this can be extended to include features of salience (cf. Machin 2011: 130–8). The materiality of the comic or picture book as a designed object, with the cover as the most salient element at first, automatically invites responses. To trigger conscious processing (System 2), defamiliarisation ruptures the flow of reading and calls for heightened attention. Miall even claims that “foregrounding is recognized by readers regardless of their literary training. Thus the literary effects created by foregrounding should be available to any reader with a basic competence in the language” (2006: 301; see also 304). Since the identification of, and reaction to such foregrounded elements can be predicted to some extent, narratologists and cognitive psychologists find it easier to work with and study feelings of this third type, as a surprise revelation is almost guaranteed to evoke some ‘appropriate’ response. They are also very useful for teachers who work with what Judith Dodge calls ‘interactive bookmarks’ (cf. 2005: 34, 41–2), which are predetermined positions in a ←146 | 147→narrative at which students are asked to complete specific tasks – often cliff-hangers in the form of suspenseful situations, moral dilemmas or the protagonist at his or her lowest point. However, they are equally relevant in the context of quiet moments and details that are easily overlooked, but may warrant slowing down and noticing, which can be done by ‘placing’ an interactive bookmark.

Instead of working with idiosyncratic reader responses that may occur at any point during the transaction with the text, which is a staple of teaching literature in the classroom, aesthetic feelings allow scientists to take control over empirical research settings. A close tie between textual structures and anticipated outcomes limits the scope of such studies which, in turn, generates more significant data. For precisely these reasons, psychological research into reading responses has to be taken with a grain of salt. Even in the published articles by some of the leading experts in the field, such as Raymond A. Mar and Keith Oatley, we find caveats such as the following: “Numerous complications plague the measurement of both reading habits and social abilities, and these issues have been highlighted by the current study” (Mar et al. 2006: 705). One of the reassuring discoveries for an academic who feels more at home in the humanities is that empirical research – at least in this context – seems to be troubled in so many fundamental ways that any blanket judgement about the inherent superiority of the latter is untenable. Literary critics, such as Suzanne Keen, Marco Caracciolo or Howard Sklar, are worried about the implicit assumptions these studies rely on, from the literary theories to the standardised tests that supposedly measure empathy or pro-social behaviour in a reliable manner. Since it is impossible to discuss individual studies at great length, I highlight general shortcomings.

The most glaring limitation is the fact that psychology professors use their own undergraduate students as test subjects, which means that a very specific group of people represents the general reading public. The second problem concerns the complexity of the reading process itself. This is how psychologists describe the room for improvement: “In order to get an accurate measurement of actual reading behaviours, an experience-sampling method (e.g., daily diaries or occasional promptings by a digital recorder) would probably be ideal for future research. These methods, however, are quite time and resource-intensive, relying upon the long-term participation of motivated individuals” (Mar et al. 2006: 706). They acknowledge – somewhat implicitly – that the circumstances under which they have to conduct their research are not conducive to accurate measurements of natural reading behaviours, which would require a lot more funding and a long-term commitment by test subjects. But even then there would always remain the caveat that readers only report those reactions that they are willing to share, that they can remember or that are explicitly asked for by the test. ←147 | 148→This takes us to the third limitation, which is the artificial setting: “Tests which show how people read sentences in laboratories […] may not reveal how people read real texts in real situations” (Emmott 2004: 92). The fourth problem is the ‘experiencing-sampling’ that Mar and his colleagues engage in: there is no conceivable manner in which the actual experience of reading could be recorded, so empirical research has to drastically reduce the complexity: “Naturally, all reader response testing involves some form of ‘manipulation’ in the sense that some aspect of the reading process or the text has been isolated for examination” (Sklar 2013: 108; see also Emmott 2004: 16, 74). Furthermore, the individual responses that are officially recorded are always reactions to specific predetermined textual features. Marisa Bortolussi and Peter Dixon provide some criteria for their selection: “At the heart of the successful development of psychonarratology is the identification of textual features. Here we present some criteria for what a valuable textual feature should be. We suggest that features should be objective, precise, stable, relevant, and tractable” (2003: 38–40). The value of these features, as Bortolussi and Dixon suggest, has more to do with the set-up of the study and the translation of individual reading experiences into quantifiable averages (cf. Caracciolo & Van Duuren 2015: 528) than their relevance to the work itself. The texts have to be very short, so that students can read them in one sitting. Many of them have been specifically written or at least adapted to work in this context. Psychologists even have a special term for these research-compatible narratives, which is “textoids” (Mar & Oatley 2008: 187; see also Emmott 2004: 16). Even with genuine literary texts, the selection process is rarely made transparent, although the whole point of the experiment is often to test how great works of literature have an impact on readers: “literary value seems to be equated with the researchers’ own assumptions and interests” (Caracciolo & Van Duuren 2015: 528). In this context, Mar and Oatley reveal a rather peculiar taste in contemporary fiction:

In more contemporary times, only a unique set of individuals succeed in producing and publishing public, crafted literary narratives. These authors are experts in understanding human psychology and behavior and may think deeply about an issue for years. By consuming the wisdom and observations of these individuals, we may thus stand on the shoulders of giants. (2008: 182)

The fifth problem has to do with the “completed experience” (Keen 2010: 84) of having read a book: referring to the post-reading tasks that psychologists hand out to their students, Keen criticises that “these questionnaires are necessarily retrospective and tell us nothing directly about either the experience or effects of reading” (2010: 85). There are some attempts to record readers’ immediate ←148 | 149→responses through ‘think-alouds’ (cf. Miall 2006: 304–5), but this is also somewhat limiting as readers may only verbalise what has reached a level of clarity and can be meaningfully communicated. Referring to a study by Kuiken et al. (2004) Marco Caracciolo and Thom Van Duuren comment in the following manner on this issue:

… it may be wondered whether the object of this study is the reading experience per se (what we may label the ‘online’ reading experience), readers’ posthoc (‘offline’) reflections on that experience, or perhaps both at the same time. […] The self-modifying feelings examined by the researchers emerge in readers’ ‘offline’ commentaries, but the source of these feelings remains unclear: is it reading the text, or is it rather the task of commenting on one of the passages? (2015: 530; see also Emmott 2004: 69)

And the ‘results’ of all these tests depend on the “inference-making of the researcher” (Emmott 2004: 95), as there is neither direct access to the cognitive processes themselves, nor an easy way to compare the more reader-oriented set-ups of think-alouds or reading diaries. This has to suffice for the moment as a precaution against a blind trust in empirical research that may be able to predict or elucidate certain trends or tendencies, but is riddled with difficulties the complexities of which I have not even touched upon.

Miall and Kuiken (1994) initially followed the path of researching reader responses to forms of stylistic foregrounding, which automatically limits the range of potential reactions. Readers may find unremarkable elements of a narrative – down to the single word – striking because of their unique biographies, predilections or cultural backgrounds. Miall and Kuiken, however, built on the premise “that stylistic features of literary texts deautomatize perception” (1994: 389). Of course, this does not happen randomly, but as an orchestrated attempt on the part of the writer to direct readers’ attention in such a way that they are more likely to recognise certain patterns that have been skilfully foregrounded in the narrative: “In literary texts […], foregrounding is structured: it tends to be both systematic and hierarchical. That is, similar features may recur, such as a pattern of assonance or a related group of metaphors, and one set of features will dominate the others” (1994: 390). This can be related to what Iser means by “strategies” and how they “organize the internal network of references, for it is these that prestructure the shape of the aesthetic object to be produced by the reader” (1980: 96).

However, contrary to Iser, who largely ignores emotions in his model, Miall and Kuiken embrace them as central to their concept of defamiliarisation: “de-familiarization evokes feelings, and feelings guide ‘refamiliarizing’ interpretative efforts” (1994: 392; see also 404). Since the flow of reading is purposefully ←149 | 150→interrupted to foreground elements that the readers should not miss, they need some time to re-orient or ‘re-familiarise’ themselves: “defamiliarization obliges the reader to slow down, allowing time for the feelings created by the alliterations and metaphors to emerge” (1994: 392). In this context, Miall and Kuiken even argue that both intratextuality and intertextuality serve the purpose of facilitating readers’ (re)orientation within narratives:

… the feelings accentuated while reading foregrounded passages sensitize the reader to other passages having similar affective connotations. Furthermore, such accentuated feelings sensitize the reader to other ‘texts’ (e.g., personal memories, world knowledge) having similar affective connotations […]. With such affectively congruent intra- and extra-textual resources, the reader ‘refamiliarizes’ or ‘thematizes’ the textual subject matter. (1994: 395)

As in Damasio’s model, emotions take centre stage and provide the vital link between different, only loosely connected experiences. When confronted with a character’s unfamiliar circumstances, readers are invited to bridge the experiential gulf by reaching out beyond the confines of their own lives. Since the autobiographical selves of readers are entangled in the narrative, readers notice that feeling tones cross over between their own memories and what the story offers them in terms of new experiences. Thus, they ‘feel’ their way into a narrative by anticipating potential developments: “Feeling-guided boundary crossing evokes personal memories and reflections in a manner that provides a framework for understanding subsequent narrative developments. In general, feeling exercises anticipatory effects by alerting us to the significance of an event that has begun to unfold” (Miall & Kuiken 2002: 227; see also Miall 2006: 304). This is the aesthetic complement or counterpart to text-based theories of foregrounding: we notice or ‘mind’ things because we care. Such a highly involved reading may lead to yet another type of emotional response, which completes Miall and Kuiken’s classification.

The fourth group, self-modifying feelings, are again triggered by a narrative text, but leave a lasting impression that may even change readers’ outlook on life. They are triggered by an intense reading experience that “has the capacity to implicate the self and deepen selfunderstanding” (Kuiken et al. 2004: 171). With the help of so-called “remembered emotions” (Miall & Kuiken 2002: 225; see also Oatley 1994: 62–3) readers flesh out narrative scenes by using personal experiences (episodic memories) to ‘get a feeling’ for a scene or situation.

This autobiographical comparison enable[s]; the reader partially to reinstate feelings from an earlier time in life and use them to understand story characters and their actions. Remembered feelings, that is, the reinstatement of feelings across similar ←150 | 151→situations, provide what we have called narrative feelings. In experiencing a fresh emotion, in contrast, readers realize something in a literary text that they have not previously experienced – or at least that they have not experienced in the form provided by the text. (2002: 226; see also Kuiken et al. 2004: 175)

This is when boundary-crossing or blending occurs and both input spaces – the personal memory and the character’s situation in a narrative – contribute to a blended space that generates new meanings and insights which, in turn, shed new light on the input spaces. This is a phenomenon that Miall and Kuiken base on Cohen’s metaphor of personal identification (cf. 1999), which is “a form of enactive reading that implicitly blends the fictional world with what readers know, believe, or feel about their own lives. […] In these cases, the reader is, we suggest, confronting personal feelings and recontextualizing them in the light of the fresh feelings evoked during reading” (2002: 238; see also Davis et al. 1996). What Miall and Kuiken describe here as a fresh feeling is an emergent structure in a successful blend (cf. Fauconnier & Turner 2003: 42). For the moment it is sufficient to understand that Miall and Kuiken extend the concept of scripts to “affective scripts” (2002: 226; see also Kuiken et al. 2004: 176), so that remembered emotions serve as an emotional guideline for a narrative situation. When fresh emotions are involved, and this is where it becomes interesting, they may partially or even completely rewrite these scripts. In other words: they modify our personal memories and the feelings we attach to them. Miall and Kuiken believe that “the experience of feelings in one situation leads to the re-experiencing of those feelings in situations that are similar” (2002: 226), yet not entirely equal, so that the emotions are never a perfect fit and an approximation at best. Since personal memories become malleable through recall, the feelings we experience during the reading process may equally affect our understanding of the past. Thus, we may re-evaluate a personal relationship or a particular memory in light of what we have just read.

Catherine Emmott stresses the roundabout “way in which a reader converts strings of words into mental representations of characters which can prompt the reader to feel considerable empathy with those characters” (1998: 176). These emotions are evoked and attributed, but not directly present in the text. This has two important consequences: first of all, the concept of what it means to identify with a character has to be clarified. Empathy in the sense of becoming other characters or feeling exactly like them may be an illusion. Either the similarities are more superficial than we would like to acknowledge or the way we have helped to flesh out the characters has made them very similar to ourselves. Secondly, and this is more significant in the context of self-modifying feelings, which is about reading oneself rather than others, the slight mismatch between ←151 | 152→the feelings we bring to bear on a story and the imagined experiences of the characters allows for a re-negotiation of our own feelings in the light of ‘similar’ experiences that characters have (cf. Miall 2006: 304). Thus, reading fiction may provide a safe environment to allow troubling memories to resurface, so that readers can explore them with the help of characters facing a similar situation. There are dozens, if not hundreds of picture books for children that address one challenge after the other: being afraid of the dark, facing up to bullies, getting braces, showing some courage, doing the right thing etc. Parents hope that reading these books with their children will lead to self-modifying feelings or – to put it more bluntly – have a practical effect. When the emotions are strong enough, this can lead to “boundary crossing” (Miall & Kuiken 2002: 227), what I call blending, which makes mutual influence more likely.

In this context, Kuiken et al. differentiate between two types of ‘self-implication’, which is their term for the entanglement of readers’ autobiographical selves in the reading process. They use the terms ‘simile’ and ‘metaphor’ to describe the extent to which readers’ personal memories become activated and implicated by a narrative: “the personal memories evoked during reading often capture similarities between aspects of a personal memory and aspects of the world of the text” (Kuiken et al. 2004: 183). In case of a simile, readers have the experience of a (strong) similarity, so that the protagonist appears to be very much like them. The stronger version is based on “metaphors of personal identification” that “depend upon an interaction between memories and world text that is not only self-implicating but also self-modifying. Enlivenment in this form is enactive” (Kuiken et al. 2004: 185). Kuiken et al. rely on Ted Cohen’s work to highlight readers’ potential for ‘acting out’ or ‘role-playing’ a character which may allow for a stronger presence of readers’ autobiographical selves in the experience of the scene and a greater likelihood of the modification of personal feelings:

the momentary state of a reader’s absorption within an author’s, narrator’s, or character’s perspective can become self-modifying when the reader metaphorically identifies with that figure. Cohen has in mind a mode of identification resembling dramatic enactment: a figure in literature may be brought to presence, as in method acting, through the embodying experience of the reader. Within this mode of reading, the embodied self is present but subsidiary within a performance that enlivens and extends, rather than merely mimics, the character’s demeanor in the world of the text. (2004: 179–80)

This is not the only theory that implies that ‘enactment’ may be a viable path to better understand both a character and oneself through a temporary blend of two personalities: “Within the moment of emerging metaphoric identification, the possibility of changing the reader’s sense of self also emerges. Within ←152 | 153→that transition, argues Cohen, there is an opening for self-modifying feelings” (2004: 180). This suggests that empathy and self-implication are two sides of the same coin, but not necessarily the same thing. As we shall see with Suzanne Keen’s criticism of empathy, feeling for and with characters does not automatically lead to real-life consequences of any kind. In this sense, Kuiken et al. may be right to keep them apart. Their claim is also much more modest than any educator could hope for, based on the results of their empirical research: “only some readers, perhaps those who become absorbed in experiential reading […] will develop a coherent and self-modifying understanding of the meaning of foregrounded passages” (Miall & Kuiken 2002: 229). The self-fulfilling prophecy is very strong here: those who identify as avid readers are very good at it and naturally empathic (cf. Mar et al. 2006: 698, 707). Those who struggle with reading find it difficult. One last thing that has to be addressed is the difference between a literal performance in front of a public audience and the kind of enactment Kuiken et al. are referring to, which takes place during readers’ very personal transactions with a text. Self-implication becomes a possibility in the second case, precisely because the text triggers a (re)negotiation of very private concerns in a safe environment.

3.3.3 Transportation

In his 1993 study Experiencing Narrative Worlds Richard Gerrig concentrates on “two metaphors that are often used to characterize experiences of narratives: readers are often described as being transported by a narrative by virtue of performing that narrative” (1998: 2). This suggests that strong cognitive involvement and empathy may not only lead to self-implication and self-modification, but also to ‘transportation’. What makes ‘transportation’ or reading is travelling a hazy concept, however, is the general vagueness of what exactly is involved. Since the heyday of reader-response criticism Gerrig (cf. 1998), Melanie C. Green and their colleagues have conducted substantial empirical research in the area of ‘transportation’, but this has not really helped to clarify the concept beyond the ‘engrossment’ and ‘immersion’ that Benton and Fox refer to (cf. 1985: 12) or that Collie and Slater praise: “The reader is eager to find out what happens as events unfold; he or she feels close to certain characters and shares their emotional responses. The language becomes ‘transparent’ – the fiction summons the whole person into its own world” (1988: 6; see also Mar & Oatley 2008: 175).

In their article “Transportation: Challenges to the Metaphor” (2015) Bortolussi and Dixon criticise that “the metaphor of transport is accepted at face ←153 | 154→value” (2015: 527) by their colleagues, which makes this “all-or-none, unitary approach” both “simplistic and misleading” (2015: 528). There is no doubt that readers report back that they feel very close to characters and right there next beside them, but what does that mean precisely? At one point, Green and her colleagues suggest that transportation is the starting point and not the end result of a successful engagement with a text:

Transportation into a narrative world may be a prerequisite for identification with fictional characters. Central to the process of identification is the adoption of a character’s thoughts, goals, emotions, and behaviors, and such vicarious experience requires the reader or viewer to leave his or her physical, social, and psychological reality behind in favor of the world of the narrative and its inhabitants. (2004: 318)

This seems logical in the context of ‘transportation’, as we first have to enter the storyworld to meet the characters, but it also illustrates the limitations of the metaphor, which draws too much attention to a literal journey. When Iser compares reading to travelling, he means that the transaction with the text is an ongoing experience, that it can be exhausting but equally rewarding at times, that we encounter different points of view that allow us to look at the world from different angles and, finally, that, what we later call ‘the journey’, is a retrospective abstraction of very immediate experiences along the way. In comparison, Gerrig and Green’s ‘transportation’ remains ambiguous, as it covers a whole spectrum of meanings that range from the literal to merely being actively engaged.

Gerrig’s ‘performance’ – the central element of transportation – resembles reader-response criticism’s active transaction with the literary work at one time, but may also refer to a close identification with specific characters, which is arguably not the same thing. Iser stresses the multiplicity of perspectives that readers have to coordinate over strongly identifying with the protagonist. This becomes obvious when Green et al. address the possibility of leading vicarious lives through role-playing and identifying with characters: “In essence, to identify with a character means seeing the character’s perspective as one’s own, to share his or her existence. Achieving such an altered state of awareness relies upon transportation into the story world” (Green et al. 2004: 319).

Overall, Gerrig’s Experiencing Narrative Worlds is a fascinating exploration of the impact of reading experiences on real life, but the two central metaphors remain underdeveloped. Gerrig states that they “serve both as shorthand expressions for what it feels like to experience narrative worlds” (1998: 2), by which he means that “a narrative serves to transport an experiencer away from the here and now” (1998: 3; see also Green et al. 2008: 513). Yet, the same applies to daydreaming or any other flow experience during which humans become lost ←154 | 155→in the activity and forget what is happening around them. If ‘transportation’ is similar to ‘flow’ (cf. Green 2004: 248; Green et al. 2008: 513–4, 532) or automatic, non-conscious, System 1 consistency-building (cf. Green et al. 2004: 315), where is the contribution of System 2, Bortolussi and Dixon ask (cf. 2015: 528–9). Green seems to suggest that the experience of flow keeps ‘transportation’ or active involvement alive, whereas conscious attention would take readers ‘out of’ the story. Bortolussi and Dixon take the opposite view, which is also problematic: “Our conclusion then is that the comprehension of literary narrative requires of necessity demanding, knowledge-driven cognitive processes and that highly transported readers must engage in a high degree of elaboration” (2015: 531).

The problem, as always, is discussing such matters without any specific context. Bortolussi and Dixon seem to have the Great Canon of English Literature and literary analysis in mind, for which a sustained flow experience would be highly unlikely. Green and Timothy C. Brock’s research, however, is focused on reading for pleasure, which in the case of ‘genre fiction’ (cf. Green & Brock 2000: 703) may indeed require little conscious attention. Here is how they describe the narrative they selected specifically for their research: “ ‘Murder at the Mall’ is a true story about a little girl, Katie, who goes to the mall with her college-age sibling. While at the mall, Katie is brutally stabbed to death by a psychiatric patient. The tragic story is moving and shocking” (2000: 703–4). I do not mean to ridicule this type of research, but it helps to put the results into perspective.

In her article “Understanding Media Enjoyment”, co-authored with Timothy Brock and Geoff Kaufman, Green specifically addresses the question of escapism and how “an enjoyable media experience […] takes individuals away from their mundane reality and into a story world” (Green et al. 2004: 311; see also 314, 317; Green & Brock 2000: 702). The concept of a ‘page-turner’ is tailor-made for the uninterrupted flow of reading, during which it would be awkward to be taken out of the experience to search for the deeper meaning. Green and Brock embrace this form of unreflected or naïve reading as it naturally occurs in real life, especially with mainstream bestsellers:

Transported readers may be less likely to disbelieve or counterargue story claims, and thus their beliefs may be influenced. Next, transportation may make narrative experience seem more like real experience. Direct experience can be a powerful means of forming attitudes […], and to the extent that narratives enable mimicry of experience, they may have greater impact than nonnarrative modes. Finally, transportation is likely to create strong feelings toward story characters; the experiences or beliefs of those characters may then have an enhanced influence on readers’ beliefs. (2000: 702; see also 703; Keen 2010: 83, 102)

←155 | 156→

Although ‘getting into’ a book or even ‘becoming lost’ in one seem desirable from an educational point of view, the complete identification with a character and the narrative’s strong appeal to emotional responses eradicate all critical distance. Green and Brock address this point again in their discussion of results, which provide “further support for the distinction between transportation and cognitive elaboration” (2000: 712; see also 718). In other words, transportation is a System 1 phenomenon. In educational settings, reading as a process has to go back and forth between the intimate communion between readers and characters, the co-construction of meaning in small groups and more analytical tasks that require rereading and a critical stance. Otherwise, students may accept a text at face value: “Individuals’ immersion in a work of literature may allow the implications of the narrative to become part of the reader’s real-life beliefs” (Green 2004: 247).

The biggest problem with transportation and ‘going further’ into the world of the story is the suggestion that readers ‘leave’ their familiar surroundings and explore a new world while their own reality fades away: “To imagine what has been stimulated by aesthetic semblance entails placing our thoughts and feelings at the disposal of an unreality, bestowing on it a semblance of reality in proportion to a reducing of our own reality” (Gerrig 1998: 21; see also 173). Yet, a successful transaction with a text allows readers to familiarise themselves with ‘the other’ and realise that it has not been all that strange in the first place and can be integrated into their current understanding of the world (cf. Sklar 2013: 12). Gerrig quotes Marie-Laure Ryan’s “principle of minimal departure” (1998: 13; see also Hogan 2003: 117; Stockwell 2002: 96; Palmer 2004: 35), which suggests that the gaps in the story world can be automatically filled precisely because the worlds are so similar. Thus, ‘transportation’ is a misleading term for a process that could equally be called ‘assimilation’ or ‘integration’. What is more, from the perspective of reader-response criticism, there is no fully realised world to discover in the first place, but only a blueprint. What I can find out there or – in the case of books – in there is already made – at least partly – from the building materials available to me. The literary work of art does not so much invent completely new worlds, but defamilarises the ones we know. When Howard Sklar states that “the intensity of readers’ emotional responses to narratives depends greatly on the proximity of the events and the situations of the characters to their lives” (2013: 20), readers do not have to travel very far to have meaningful experiences.

Research in transportation is full of trivial insights, as System 1 is automated and thus inaccessible, which means that psychologists have to depend on readers’ self-reports. Nevertheless, Green and her colleagues are undeterred in their conviction that “[t];ransportation into a media world can be measured with ←156 | 157→a 15-item self-report scale” (cf. Green 2004: 313). The most consistent results are that avid readers report strong feelings of enjoyment, transportation and identification with characters, whereas “less-skilled readers find it harder to become thoroughly immersed in narrative” (Gerrig 1998: 19). Well-written narratives are more likely to draw in readers than ‘bad’ ones: “Just as a leaky boat does a poor job of transporting people across the water, poorly constructed narratives do not help readers enter the story world” (Green et al. 2004: 320). Yet, it can get even more basic than that: “Circumstances that prevent readers from being fully immersed in a narrative world reduce media enjoyment” (Green et al. 2004: 321). Further relevant factors include personal relevance, emotional investment and prior familiarity with the topic or the narrative itself (cf. Green 2008: 515, 522).

What research in transportation phenomena does demonstrate is that the key characteristics of successful reading mutually reinforce each other: a good book is more likely to draw in readers, who are more willing to be transported and become engaged with the characters, which makes them enjoy the narrative a lot more and lets them experience something new, which they feel is meaningful to their own lives. Although this may sound trivial, there is a kernel of truth that one can take away from this: reading, like any other human pursuit, depends upon a successful entry point that is going to positively reinforce other aspects, which in turn are going to contribute to the overall feeling of getting noticeably better at something. What it takes is to find the right books for every student.

3.3.4 Empathy

Finally, we can turn to Suzanne Keen’s work on narrative empathy, which provides some much needed orientation. Like ‘genre’, ‘empathy’ has been used for such a wide range of phenomena that the term has almost lost its critical potential. Keen is willing to ask uncomfortable questions and approaches the subject matter within a much broader framework than is usually the case. She senses a “resistance to empathy [that] is cultivated by academic modes of analysis that privilege critical distance and observations about style” (2010: 73), while she sees herself as a conscious supporter of “bringing affect to the center of cognitive literary studies’ reexamination of narrative fiction” (2010: xii). That this has not always been easy for a scholar closely affiliated with narratology and gender studies, makes for a more interesting engagement.

Empathy is a prerequisite for readers’ self-implication in narratives and the self-modifying feelings that, according to Miall and Kuiken, can literally change a person’s life. The two most popular ways of describing empathy is “the capacity to take another’s perspective” (2010: 27) or “to step into a character’s ←157 | 158→shoes” (2010: 18). Although this basic idea of perspective-taking may seem straightforward and self-explanatory, there are several areas of critical contestation that make the matter more complicated: first of all, the term ‘empathy’ only entered the English language in the early twentieth century (cf. 2010: 55), but then gained so much currency that it began to replace ‘sympathy’ without completely eradicating it. The second problem concerns the goal of empathy: do we become better human beings, better readers, better mind-readers or maybe all three things in one package? This, in turn, raises the question whether ‘narrative empathy’, which is Keen’s term (cf. 2010: xxv), and real-life empathy are the same, closely related or even significantly different. Do we predominantly learn to understand ourselves better, as Miall and Kuiken seem to suggest, or others by sharing our feelings with them? The third problem is tied to the “multiplicity of reactions” (2010: 95) that are labelled as ‘empathy’, which is going to be the main focus for this section. A fourth problem concerns the question whether empathetic feelings have to be disentangled from cognition and studied separately: is empathy a spontaneous emotional reaction, part of an ongoing cognitive process that drives our meaning-making or an object of study in itself?

In his article “These Things Called Empathy” Charles Daniel Batson, a social psychologist, differentiates between eight different phenomena that have been referred to as ‘empathy’, each of them representing “a conceptually distinct, stand-alone psychological state” (2009: 3). Since Keen adopts this classification in her article “Intersectional Narratology in the Study of Narrative Empathy” (2015), it seems appropriate to start with this categorisation and then critically discuss empathy in more general terms again. I adopt Batson’s exact headlines to highlight how he distinguishes between the concepts and then I further develop them by adding cognitive (literary) theories that, I believe, fit the categories. As with the distinction between aesthetic and efferent reading, I consider it helpful to keep the eight types conceptually apart, despite the fact that there are overlaps and that some distinctions may be too specific for educational settings.

“Concept 1: Knowing Another Person’s Internal State, Including His or Her Thoughts and Feelings” (Batson 2009: 4)

This has come to be known as Theory of Mind or simply “mind-reading” (Zunshine 2006: 4) in cognitive literary studies and is mostly associated with the work of Alan Palmer (2004; 2010) and Lisa Zunshine (2006). Due to its strong affiliation with narratology, I consider it closer to ‘theory theory’ than simulation theory (concept 6), but the two are often taken together. As a thoroughly analytical approach, Theory of Mind circumvents readers’ emotional investment in the narrative by claiming that we can reach an understanding of characters’ ←158 | 159→thoughts and feelings through a narratological analysis of the text. For literary scholars interested in cognitive studies, this has been a softer transition than fully embracing reader-response criticism or enactivism. Still, Theory of Mind breaks with classical narratology in significant ways, mainly by placing the characters and their social interactions in the centre of attention. This leads Palmer to denounce the traditional plot-fixation of narrative analysis:

It is difficult to combine subjects into a plot structure without compromising their subjectivity because they would then become simply elements in a narrative framework. This danger can be avoided if the idea of plot includes some notion of the multiplicity of characters’ discourses and therefore becomes a more organic and flexible concept than the traditional approach. (2004: 156)

He redefines the actions or events of narratives as experiences of the main characters and thus follows Monika Fludernik’s concept of experientiality (cf. 2004: 31–2). Since we can only understand characters through their interactions with others in specific contexts, Palmer emphasises “the process of reading and not the end product. The embedded narrative approach is primarily an attempt to explore fully the workings of dense and complex fictional texts. This is the process. The end products are the various purposes to which these explorations might be put” (2004: 21). He regrets “the fact that narratology has created clear boundaries between various aspects of fictional minds, even though the fictional texts themselves show that these boundaries are not clear at all” (2004: 28). Characterisation, focalisation and the representation of consciousness are seen as distinct areas of narratology (cf. 2004: 43), despite the fact that they are all key sources of information on characters. This general reorientation in cognitive narrative studies towards the centrality of characters is significant, even if it represents just a first step in the direction of embodied cognition.

However, looking at the details of Palmer’s approach, it becomes obvious that it is indebted to schema theory and classical narratology in other respects. He claims that “we assemble […] an embedded narrative”, that encompasses “the whole of a character’s various perceptual and conceptual viewpoints, ideological worldviews, and plans for the future considered as an individual narrative that is embedded in the whole fictional text” (2004: 15; see also 121). The necessary information can all be found in the text, so readers have to be trained to extract it in its entirety:

The mental events, processes, and states that distinguish actions from mere doings are crucial to the concept of embedded narratives. A description by a narrator of a character’s action is a description of the development of that character’s embedded narrative. The reasons, motives, intentions, purposes and so on behind the action may be ←159 | 160→explicitly specified by the narrator, they may be implicit but understood by the reader, or they may remain mysterious. However, they are always there in the storyworld. The core of the embedded narrative approach is the systematic analysis of the structure of mental events that lies behind the decisions that lead to actions and, specifically, of how this is presented in the discourse by the narrator. (2004: 122)

The impact of classical narratology is very strong here, as reading is conceptualised as “the systematic analysis of the structure of mental events” (2004: 122), for which Palmer replaces a traditional narratological category with his own idea of ‘embedded narrative’. This approach may be more relevant to professional narratologists, like Palmer himself, but not to the general reading public, who do not “strongly prefer to read a text for maximum cognitive payoff” (2004: 176). If teachers prefer a “narratology in action” (1992: 51), as Michael Benton suggests, it has to help students to transact successfully with texts and acknowledge their individual cognitive processes.

‘Theory of Mind’, as the term suggests, operates with a form of ‘folk psychology’ – “everyday assumptions about the workings of the human mind” (Nünning 2014: 134) – and a ‘personality model’ (cf. Nünning 2014: 272–3) that allows humans to read other people’s minds in real-life encounters, but also characters’ behaviour in fiction. Zunshine claims that “our ability to explain people’s behavior in terms of their thoughts, feelings, beliefs, and desires” (2006: 6) is trained on a daily basis, as the attribution of “states of mind is the default way by which we construct and navigate our social environment, incorrect though our attributions frequently are” (2006: 6). In fact, we are so used to it, that Zunshine describes mind-reading as a System 1 operation: “our tendency to interpret observed behavior in terms of underlying mental states […] seems to be so effortless and automatic (in a sense that we are not even conscious of engaging in any particular act of ‘interpretation’) because our evolved cognitive architecture ‘prods’ us toward learning and practicing mindreading daily from the beginning of awareness” (2006: 7; see also 16, 85). Like all System 1 operations, mind-reading only has to be “ ‘good enough’ for our everyday functioning: however imperfect and fallible, they still get us through yet another day of social interactions” (2006: 59).

In this context, literature can even be seen as a training ground for this “ability of human beings to make sense of the actions, intentions and thoughts of others”, as Vera Nünning explains (2014: 131), precisely because the literary treatment makes fictional minds and their various interactions more coherent and accessible, in some cases inviting us – through foregrounding – to slow down and pay closer attention. In contrast to real life, where we often rely on System 1 for quick orientation, literary texts provide readers with unique opportunities to dedicate ←160 | 161→all their mental resources to an exploration of other people’s minds. This ties narrative competence – or “narrative intelligence” in Herman’s words (2002: 1) – directly to the skill of mind-reading (cf. Nünning 2014: 133; see also 136, 150). Vera Nünning offers a useful summary of how this is conceptualised:

Narratives structure, interpret and communicate knowledge about the human mind, about human communication, and human behaviour. They are based on and contribute to the interpretability of human experience. They can also serve to explain and disseminate beliefs about the way human minds work in specific interactive situations. Stories thus often popularise knowledge about theory of mind, but they do so not by way of disseminating abstract principles. Instead, they present specific acts of reflecting, communicating and behaving, and delineate the dynamic interactions between several characters. By showing human beings in specific (interactive) situations, they also enhance readers’ understanding of the nature and scope of interactive encounters and of human communication. (2014: 150–1)

Relying on readers’ highly advanced mind-reading skills, writers may take the opposite route and dispense with the usual assortment of ‘empathetic narrative techniques’ (cf. Keen 2010: 92–9). The twentieth century saw quite a few experiments of this type:

Hemingway could afford such a deliberate, and highly elaborate, in its own way, undertelling for the same reason that [, in Mrs. Dalloway,] Woolf could afford to let Peter’s trembling “speak for itself”: our evolved cognitive tendency to assume that there must be a mental stance behind each physical action and our striving to represent to ourselves that possible mental stance even when the author has left us with the absolute minimum of necessary cues for constructing such a representation. (Zunshine 2006: 23)

One point that has not been stressed enough is that, from a Theory of Mind point of view, and even more so in the enactivist paradigm, humans and literary characters do not have rich inner lives, but rich social and cognitive interactions with their environments. The need to read other people’s minds only ever arises during social interactions, which means that the attribution of a state of mind is not so much based on guessing than on the study of circumstantial evidence in the form of speech, behaviour, gestures, body postures, facial expressions, proxemics etc.

Just as in real life the individual constructs the minds of others from their behavior and speech, so the reader infers the workings of fictional minds and sees these minds in action from observation of characters’ behavior and speech. In one sense […] we are invisible to each other. But in another sense the workings of our minds are perfectly visible to others in our actions, and the workings of fictional minds are perfectly visible to readers from characters’ actions. Most novels contain a wide variety of evidence on ←161 | 162→which readers base their conjectures, hypotheses, and opinions about fictional minds. (Palmer 2004: 11)

This leads Palmer to the “position that meaning is not inner, mysterious, private, and psychological, but outer, evident, public, and behavioral” (2004: 142). Mind-reading, in this sense, is a form of “social cognition” (Nünning 2014: 133) that can be fruitfully explored in contexts where humans lack this skill, which is the case with autism and Asperger’s. The great success of Mark Haddon’s The Curious Incident of the Dog in the Night-Time can be understood in the context of foregrounding a complete lack of Theory of Mind, which is unusual for a general reading public that takes this ability for granted. In Fictional Minds, Palmer proclaims that the constructions of the minds of fictional characters by narrators and readers are central to our understanding of how novels work because, in essence, narrative is the description of fictional mental functioning” (2004: 12). While Palmer’s reorientation of narratology in view of readers’ engagement with characters is an important first step, his overreliance on exhaustive analysis makes it difficult to reconcile this approach with Miall and Kuiken’s readerly feelings.

“Concept 2: Adopting the Posture or Matching the Neural Responses of an Observed Other” (Batson 2009: 4–5)

This idea has also been conceptualised as ‘motor mimicry’ or ‘facial empathy’ (cf. Batson 2009: 4) and was recently substantiated by the discovery of mirror neurons (cf. Gallese 2009). In this context it can be understood as a largely automated, physiological process that can even be observed in infants. Batson refers to some attempts to build a whole theory of empathy based on the close ties between neural response matching and motor mimicry (cf. Batson 2009: 5), which he quickly dismisses:

Perceptual neural representations do not always and automatically lead to feelings, whether matched or unmatched. And at a motor level, neither humans nor other species mimic all actions of others. To find oneself tensing and twisting when watching someone balance on a tightrope is a familiar experience; it is hard to resist. Yet we may watch someone file papers with little inclination to mimic the action. Something more than automatic mimicry must be involved to select those actions that are mimicked and those that are not. (2009: 5)

Amongst others, Vittorio Gallese and Michele Guerra have made efforts to tie mirror neurons to embodied cognition and to conceptualise new approaches to empathy in films (cf. Badt 2013), but such endeavours remain largely speculative for the moment. A second line of argument concerns the deliberate use of ←162 | 163→mirroring facial expressions as a “higher-order communicative function” (Batson 2009: 5), usually to signal to another person that one understands how they feel.

“Concept 3: Coming to Feel as Another Person Feels” (Batson 2009: 5–6)

This may occur in the form of emotional contagion – ‘catching’ a group’s dominant response to an experience, e.g. as a member of an audience – or as empathy in a narrower sense, ‘matching’ emotions that are close to or exactly like those of another person. I have already discussed this second meaning in the context of Miall and Kuiken’s self-modifying feelings (cf. 2002: 232) or Gerrig and Green’s transportation. In contrast to concept 1, “coming to feel as another person feels” usually circumvents mind-reading by either being swept away by a mood or wave of emotions or employing affective scripts (remembered emotions) to close the distance between readers and characters, which in turn may lead to boundary-crossing and a feeling of identification, even if there can never be a perfect match. In Werner Delanoy’s example of watching Dead Poets Society with a group of university students we have already encountered a situation in which viewers experienced transportation, reported a complete identification with the protagonists, and showed signs of both self-implication and self-modifying feelings in the form of repercussions on the students’ beliefs (cf. Delanoy 1996: 65–6). While ‘transportation’ may be the only way to have students fall in love with narratives, it is not the easiest reading response to work with in the classroom. Over-identification leads to blindness in other regards and negates the obvious distance between readers’ diverse selves and the circumstances depicted in a narrative. As co-creators of the story, students may project too much of themselves onto the characters and fail to notice subtleties or cultural references that would otherwise defamiliarise the content. This plays a crucial role in the context of concept 4, where this discussion is continued.

Mimicry (concept 2) may play a role in coming to feel as another person, by triggering an emotional response, which can then be reflected back to the sender – in the form of a smile, for example. This is easier with basic emotions or what Keen calls ‘primitive empathy’: “Primitive empathy, or the phenomenon of spontaneously matching feelings, suggests that human beings are basically similar to one another, with a limited range of variations” (2010: 15). There has been substantial research on universal or basic emotions (cf. Ekman 2007), which has shown that they are essentially the same in all cultures, but heavily influenced by local conventions of how they may be acknowledged or openly displayed. These “display rules” (Ekman 2007: 4) become obvious when conducting research in cultures that frown upon the public exhibition of strong and/or negative emotions. However, the facial expressions that correspond to the eight basic ←163 | 164→emotions can be read by all humans without specific cultural training, with some variation in recognition: “The results were very clear-cut for happiness, anger, disgust, and sadness” (Ekman 2007: 10).

“Concept 4: Intuiting or Projecting Oneself into Another’s Situation” (Batson 2009: 6)

This denotes an attempt to imagine oneself under different circumstances and may be called ‘aesthetic projection’ or ‘aesthetic empathy’. It is closely tied to the concept of Einfühlung, the German source of ‘empathy’. This could be a special gift of writers who are said to be more empathetic than the average population (cf. Keen 2010: 121, 123, 130–1). It involves an exercise of one’s own imagination and engages the mind in daydreaming and make-belief. In many ways it is an exercise in creative writing and very popular as an activity for the literary classroom (e.g. diary entry; speech and thought balloons). In contrast to concepts 1 and 3, ‘projection’ relies more on creativity and playful exploration to compensate for a lack of available information or shared experiences. Batson offers an interesting take on ‘projection’ that is relevant to educational settings:

… when the state of the other is obvious because of what has happened or been said, intuition or projection is probably unnecessary. And when the other’s state is not obvious, intuition or projection runs the risk of imposing an interpretation of the other’s state that is inaccurate, especially if one does not have a precise understanding of relevant differences between oneself and the other. (2009: 10)

This problem of ‘empathic inaccuracy’ (cf. Keen 2010: xiii, xxiv, 130, 139–40), which in this case means a false sense of understanding by imposing one’s own theory of the world onto a character or real person, is a major concern of Keen’s: “What if our empathy with others is only egoism, recognition of the self, painted over the other’s true experience?” (2010: 130).

As an activity for the classroom, projection is essential to inter/transcultural learning (cf. Freitag-Hild 2010: 113–14). In the pure form of Batson’s concept 4 it would be a tightrope walk between a potentially boring, but accurate regurgitation of what has been established in a narrative and a more exciting extrapolation that may lead students in directions that are hard to reconcile with the text. In the first case, the purpose of the activity would be lost and, in the second case, potential misreadings could be encouraged. Some creative tasks for the classroom, especially in the area of roleplays, can take the explorative aspect of projection quite far. Yet, there are two important safeguards in educational settings that prevent such activities from losing their undeniable worth. The first is that students are usually asked to gather enough information about characters, usually via a rereading task, and to compare their findings. Secondly, these learner ←164 | 165→texts are then discussed, scrutinised, peer reviewed or simply commented upon by the teacher, especially when they are highly explorative. Aesthetic reading relies on meaning-making under the guidance of the text, so projection activities also have to lead from subjective impressions to the co-construction of meaning. In the case of enactment, it is necessary to discuss the performance with the whole class afterwards. Like all performances, roleplays are interpretations of the narrative, whose purpose is to provide a contribution to an ongoing discussion of a literary text. Projection tasks can be very quickly created, but they do require some considerations concerning the right balance between students’ knowledge and speculation, but also the purpose of such learner texts for the immediate and ongoing exploration of a narrative.

“Concept 5: Imagining How Another Is Thinking and Feeling” (Batson 2009: 7)

This is often referred to as ‘perspective-taking’, which represents a desirable and somewhat idealised reading position to be in, as it combines the analytical distance of concept 1 with the imaginative, empathic approximation that concept 4 endorses. All too often, perspective-taking is confused with the idea that “we take on the emotional experience of another as our own” (Sklar 2013: 24), which is not even the case with boundary-crossing in concept 3. The narrative, often through the protagonist’s experiences, functions as a catalyst that can exercise transformative power over the reader, but not by filling an empty container with a new emotion that was not there before (cf. Oatley 1994: 69). The problem for the teacher of literature is to find the right balance between a narratological study of characters’ thoughts and feelings and a complete immersion. Students may be in need of a gentle reorientation, depending on their current stage of transaction with the narrative text. Though I disagree with Sklar’s notion that readers can literally become the characters, he still has a point concerning the dangers of too much subjectivity:

These components of empathic experience suggest that empathy for a fictional character essentially places readers inside the experience – and particularly the emotional experience – of that character. Our immersion in that experience, furthermore, momentarily may impede, or temporarily suspend, our capacity to form judgments about that character, since we may, as it were, become too close to view the character’s reality objectively. In other words, since readers’ empathy may make it difficult for them to judge a character from the outside, they will tend to judge the situations that occur in the narrative from the character’s point of view. (2013: 48; see also 14, 53)

Suzanne Keen argues that “character identification […] remains the single most important facet of response to fiction articulated by middlebrow readers” (2010: 60; see also 68). This is especially problematic with autobiographical texts ←165 | 166→that invite a strong identification with the main character. Eventually, teachers have to drive a gentle wedge between readers and texts to establish a more objective point of view. The “manipulation of distance”, which may lead to an experience of “collapsed aesthetic distance” (Sklar 2013: 49), encourages a false sense of identification. Reader-response critics, like Iser, would also agree that a narrative consists of more than one perspective and that readers have to navigate between them. The problem is often over-identification, as we have already seen, and requires some active, task-based disentanglement.

‘Perspective-taking’, in this sense, can be likened to a half-subjective ‘over-the-shoulder shot’, a placement of the reader/viewer right next to a character without direct identification. This optical metaphor tends to obscure other important aspects, such as the central role of feelings, but it is good enough to capture this relative position. Since all eight categories are not pure types and may occur intermittently throughout a reading, it is futile to argue in favour of one to the detriment of the others. A general preference for ‘perspective-taking’ in educational contexts is its combination of two very popular approaches to empathy (1 & 4), which allows for rational analysis and the involvement of readers’ imagination. Since concept 3 is also dear to avid readers, but harder to work with in the classroom, it should still be possible to devise activities and ask questions that are not meant to be shared and discussed in class afterwards, but may support students’ transaction with the text. Britta Freitag-Hild’s category of “Interpretations- und Einfühlungsaufgaben” (2010: 113–14) mirrors the hybrid nature of ‘perspective-taking’: it combines Theory of Mind with aesthetic projection, narratology with creative writing, an efferent with an aesthetic stance. This has more to do with the belief that both are necessary to understand literary characters than with any sense of homogeneity within this group of activities.

“Concept 6: Imagining How One Would Think and Feel in the Other’s Place” (Batson 2009: 7)

In contrast to concept 5, which has an analytical/narratological component through its affiliation with Theory of Mind (‘theory theory’), simulation theory explores other people’s entanglements, thoughts and feelings through role-taking. Palmer first describes the concept as an approach that a method actor would take in the theatre, but then highlights its indebtedness to a computational model of cognition: “Simulation is not imagining me in that situation: it is imagining being the other in that situation. It means pretending to have the same initial desires, beliefs, and other mental states as the other person. We feed these into our inferential cognitive mechanism that then generates further mental states” (2004: 143). Simulationists believe that we can run invented scenarios ←166 | 167→in our brain to determine their outcome, which helps us in the areas of risk-management and making important decisions. The same ‘virtual-reality software’ or “planning processor” (Oatley 1999: 444), they claim, can be employed to step into the shoes of other human beings and fictional characters: “Whereas computer simulations run on computers, literary simulations (drama, short-stories, novels) run on minds, in the imagination, or like a kind of guided dream” (Oatley 1999: 441). While this already sounds like a bold claim, simulationists even propose that we recreate the entire story world and run the scenarios within this setting, backwards and forwards (cf. Oatley 1999: 444), which allows us to build consistency and predict future actions: “A literary simulation […] models objects, their attributes, and the interactions among the objects in the story world. Here, the objects almost invariably include human agents. The simulation works if a reader or spectator can get the whole thing to run – to imagine the story world with its people, and to become absorbed in it” (Oatley 1999: 441). This reveals a certain affinity to transportation, but the source of the experience is said to be a computation that we normally use to predict social scenarios. Varying the input data slightly, we can “mentally ‘try out’ different versions” (Nünning 2014: 249), which can then be appraised and ranked in terms of desirability.

Keith Oatley, one of the leading proponents of this theory, suggests that this represents the off-line mode of an interactive device that we rely on in all social situations: “A convincing theory is that consciousness is not so much a mechanism for deciding what to do immediately, but is a kind of simulation in which we relate our knowledge and memories of other people and ourselves to the current social situation and to possibilities for future social action” (2016: 624). This leads him to claim that fiction is also a kind of simulation, as authors use exactly the same procedure to conjure up stories (cf. 2016). However, since works of art go beyond simple mimesis and create new (social) worlds, they can also be seen as metaphors extended in time (cf. 2016: 618) that come with “directions to the reader about how to run the simulation” (1999: 443), which corresponds somewhat to Iser’s strategies. Although Iser may seem overtly vague about the exact process of consistency-building or meaning-making, he at least proposes a tentative theory. Oatley and his colleagues are surprisingly taciturn about how this ‘planning processor’ is supposed to work.

In contrast to the personality theory and folk psychology of concept 1 (often called ‘theory theory’), simulation theory proposes that we do not analyse characters, but understand them based on the outcome of our internal ‘stage plays’. In this sense, simulation theory is more enactive and process-oriented and thus closer to concept 4 than to any of the other dimensions of empathy presented here. Despite these conceptual differences, ‘theory theory’ ←167 | 168→and ‘simulation theory’ are understood as two strands of Theory of Mind (cf. Nünning 2014: 136–7; Batson 2009: 9), which makes matters more complicated than necessary. I treat them as separate approaches here (concepts 1 & 6) for the reasons presented above and choose to associate Palmer and Zunshine’s Theory of Mind with analysis and narratology, while I associate the role-taking aspects of simulation theory with concept 6.

Concept 4, aesthetic projection, may seem to be the same thing as simulation, but it is not based on cognitive or psychological theories and simply relies on Einfühlung and the imagination (cf. Batson 2009: 7). In both cases readers have to work with a script or blueprint that is handed to them, but they can only rely on their own resources to make sense of it and ‘perform’ it: “The process of simulation is therefore dependent on and intertwined with the projections, value judgements and anticipations of readers” (Nünning 2014: 253). Another way of looking at the difference between concepts 4 and 6 is in terms of cognitive play (cf. Nünning 2014: 23, 73, 86, 193–4) versus the more ‘technical’ understanding of simulation: “simulation theory, like pretense theory, relies on a relatively ‘mechanical’ metaphorical construct to explain the reader’s mental activity while reading. With make-believe [concept 4], the reference is to the structure and processes of children’s play. In simulation theory, more often than not, the reader’s mental activity is compared with that of a computer” (Sklar 2013: 17).

Despite this technically conceived simulator or ‘holodeck’, Oatley and his colleagues insist on the importance of role-playing and vicarious experience (cf. Nünning 2014: 36) that lets us imaginatively broaden our horizon and the range of our experiences. Simulationists believe in the whole package that reading provides: that we become better human beings, better readers and better mind-readers (cf. Mar & Oatley 2008: 180). They argue that this transformational power is due to the unique quality of literariness (cf. Mar & Oatley 2008: 182), so that the way “the events in a fictional story are carefully selected, foregrounded and manipulated adds to the value of fiction” (Nünning 2014: 39). In contrast to perspective-taking (concept 5), role-taking tends to maintain a clearer difference between actor and role. Readers have to simulate all the characters, not just the charming and relatable hero(ine), in the same way that a method actor has to transform into a mass-murderer when needed. Since easy identification is not always an option, ‘adopting the goals of a protagonist’ (cf. Oatley 1994: 69; Palmer 2004: 129) is considered a vital aspect of simulation theory, as an understanding of a character’s past, present situation and motivations may be sufficient for some form of identification and emotional involvement. This is how Mar and his colleagues explain this aspect of role-taking:

←168 | 169→

One need only attend a Shakespearean play to observe that understanding and keeping track of the motivations, intentions, and beliefs of characters is paramount for narrative comprehension. Narratives are fundamentally social in nature in that almost all stories concern relationships between people; understanding stories thus entails an understanding of people, and how their goals, beliefs and emotions interact with their behaviours. (2006: 696)

Here, a strong affinity to ‘theory theory’ (concept 1) comes to the fore, which may explain the umbrella term of Theory of Mind for both approaches.

Simulation theory suffers from two fundamental problems: an underdeveloped concept of a ‘planning processor’, whose ontological status and base of operations are not sufficiently explained, and a tendency to collect various cognitive theories under the umbrella term of ‘simulation’. The best definition seems to be the notion that humans can run ‘what-if’ scenarios, for which the processor has to be fed with factual information taken from the narrative. At the same time, ‘running the simulation’ can lead to a whole spectrum of responses and developments, including a strong identification with a character. As with drama techniques in the classroom, which are the closest match in terms of task design, the results largely depend on the circumstances under which the simulation is performed.

“Concept 7: Feeling Distress at Witnessing Another Person’s Suffering” (Batson 2009: 7–8)

One of the prevailing notions within the humanities is that “empathic emotion motivates altruistic action” (Keen 2010: vii; see also Nünning 2014: 183), which is mirrored in the pop-cultural shorthand of depicting villains as sociopathic monsters or machines that are empathically impaired (cf. 2010: 9). One matter that is sometimes overlooked in this context is ‘empathic distress’ (cf. 2010: 19), which means that witnessing the suffering of others could lead to negative reactions, like anxiety, unease or a strong desire not to be confronted with the subject matter any longer. By attempting to implicate students emotionally and personally in what teachers consider highly appropriate narratives, they may still run the risk of disturbing some of them. Instead of feeling with or for the characters, readers may end up being distressed by the brutality of the depicted world. Instead of opening up and reaching out, they may retreat and block out the elements they cannot handle, which may lead to a rejection of the whole narrative. As Suzanne Keen shows in reference to Suzanne Collins’s The Hunger Games trilogy (cf. 2015: 132–3), this is not a constructed case.

Students may also feel uncomfortable or self-conscious when they are confronted with a social reality that they or their families are not willing to face. ←169 | 170→In this case, the truth may be equally hard to handle. From Sklar’s point of view, taking students out of their comfort zone and confronting them with contexts that may not be so easily digested is all part of teaching and often a necessity to promote pro-social behaviour (cf. 2013: 40). There is no simple answer to the question which texts are appropriate for the classroom. Maia Kobabe’s Gender Queer: A Memoir is a fascinating exploration of eir teenage self’s queer identity, but e depicts growing up to be a woman as a nightmare (cf. 2019: 35–7, 127–9, 220–1). Without the necessary framing, sufficiently mature students and a lot of discussions, this book has the potential to produce the same reaction as in Maia’s aunt Shari: a lingering doubt whether this consistent devaluation of everything feminine is not a misogynistic stance on some level (cf. 2019: 195).

“Concept 8: Feeling for Another Person Who Is Suffering” (Batson 2009: 8)

While empathy is associated with the concept of feeling with another person or character, pity, compassion or sympathy are feelings for others and are more directly associated with ethics and pro-social behaviour. In this case, the suffering of another human being leads directly to the wish to translate one’s reactions into real-life consequences. The interesting thing about sympathy is that it does not require perspective-taking at all: “readers need not empathize with a character in order to feel sympathy with him or her. […] we do not have to understand a character to sympathize. The fact that we can recognize suffering without necessarily having experienced the particular suffering of a given individual (or character) effectively provides us with ways of identifying ourselves with him or her” (Sklar 2013: 53; see also Caracciolo 2014: 130). This is even supported by psychological research:

To feel for another, one must think one knows the other’s internal state (concept 1) because feeling for is based on a perception of the other’s welfare (e.g., that your friend is hurt and afraid). To feel for someone does not, however, require that this perception be accurate. It does not even require that this perception match the other’s perception of his or her internal state, which is often the standard used in research to define empathic accuracy … (Batson 2009: 10; see also Mar et al. 2011: 824)

When charities run an ad on TV that shows starving African children, victims of a hurricane or survivors of a flood in Asia, most of the potential donators do not have any idea what it feels like to be a survivor of such catastrophes. Therefore, Sklar defines “four general components” (2013: 35; see also 53–9) that are required to engender sympathetic responses without having to rely on more involved forms of empathy: an awareness of suffering and ways of alleviating the situation; a judgement that this suffering is unfair and undeserved; sympathetic distress, by which he means feeling uncomfortable on behalf of the sufferer; and ←170 | 171→a desire to help. All of these neither require an identification with the afflicted people, nor an actual understanding of how they feel. Although Sklar’s strict separation of empathy and sympathy may help him to highlight the pro-social orientation that comes with sympathy and deliberation in opposition to Green’s transportation, which is essentially escapism, he concedes that many others see a continuum, or even empathy as a prerequisite for sympathy (cf. 2013: 25). This also explains why Batson lists sympathy as the eighth variety of empathy. What they all share are different forms and conceptualisations of how the gulf between two individuals can be emotionally bridged.

Keen defines three distances at which narrative texts try to engage their readers: when we are asked to identify with members of our own in-group, which could be based on nationality, ethnicity, gender, religion etc. or combinations thereof, she speaks of “bounded strategic empathy” (2010: xiv), which is supposed to be the easiest. “Ambassadorial strategic empathy” (2010: xiv) refers to an attempt to win over one particular group of people for a specific purpose, such as charity work or political reform. “Broadcast strategic empathy” (2010: xiv) involves reaching out to every potential reader by appealing to a shared humanity. These types can and do appear in combination. John Lewis, Andrew Aydin and Nate Powell’s March, a trilogy of auto/biographical comics about the Civil Rights Movement in the United States, addresses the African-American community as an in-group, invites young readers to become political activists (cf. 2013: 3) and appeals to everyone in terms of universally relatable experiences. Due to the unique qualities of literature, texts can even retarget readers’ empathy from the most general (shared humanity/human rights) to more specific concerns (The Civil Rights Movement in the United States) to single lives (John Lewis). The entry point can be different for each reader, but the narrative may manage to extend readers’ empathy to more specific concerns.

Although sympathy cannot be tied to specific genres, there is an undeniable link between victimisation or – more generally – the unfair treatment of characters (cf. Sklar 2013: 54) and many autobiographical and/or political texts that strongly encourage readers’ responses in the form of sympathy: “Empathetic anger and an empathetic sense of injustice can each lead to personal, social, and ideological responses based on understandings of unfairness or evocation of righteous indignation on behalf of victims” (Keen 2010: 18–19). Sklar argues that ‘broadcast strategic empathy’, when characters are significantly different from the readers and a shared humanity is all that ties them together, is often sufficient to appeal to readers’ sympathy. The title of his study, The Art of Sympathy in Fiction: Forms of Ethical and Emotional Persuasion, is already a strong signal that ←171 | 172→he considers ‘empathy’ or ‘perspective-taking’ as the weaker form when it comes to pro-social behaviour:

… one of the aims of schools ought to be to widen students’ sense of their “we group” in order to develop their capacity to feel compassion for people further removed from their own experience. While stories alone may not be able to foster this capacity, they can “persuade” readers to reevaluate and even to feel sympathy for those clearly, even radically, outside the boundaries of their “we groups.” (2013: 40)

While it is often suggested that narratives should be chosen that appeal to the students’ interests and facilitate a quick identification with the protagonist, Sklar invites teachers to select more challenging texts that may keep students at a critical distance from the characters, but allow for a more considerate sympathetic response than the easy path to empathy with protagonists that are too close to home: “I believe that emotional response generally, and sympathy particularly, can be usefully divided into notions of sensation and deliberation – in other words, what we feel and what we think” (2013: 29). Pro-social behaviour is a conscious decision to become active in real life and clearly needs the kind of deliberation that may be missing in the case of identification: “for emotions such as sympathy to possess ethical value, they must involve deliberation on the content of the experience itself” (2013: 32; see also Oatley 1994: 446; Keen 2010: 27–8). This returns us to Damasio’s argument that emotion and evaluation cannot be separated. For Sklar, emotions are not enough: they have to become appraisals that need to be tied to specific contexts. Stirring strong emotions in readers or viewers is often a prerogative of genres that are the least likely to have political undertones in Iser’s sense.

Although we have had a look at how empathy is conceptualised and discussed in different theories and contexts, what is still missing is a politics of the concept. Sklar positions sympathy explicitly against the widespread embrace of empathy as the more desirable reader response to narrative fiction. Since he only operates with two terms, we need to clarify what he means. Sklar seems to associate empathy with transportation (concept 3) or maybe even perspective-taking (concept 5) in case the aim is identification and not critical distance. Instead, he proposes a form of sympathy that combines emotional responses to a text with a political agenda. Otherwise, reading would lead to escapism or a self-congratulatory, ‘virtual’ engagement with a text instead of a pro-social orientation that has real-life consequences.

Keen was first in raising doubts about some of the exuberant claims that “[e];mpathy offers an almost magical guarantee of fiction’s worthiness” (2010: 62). She explicitly opposes “the empathy-altruism hypothesis” (2010: vii) by stating ←172 | 173→that “the case of altruism stemming from novel reading [is] inconclusive at best and nearly always exaggerated in favor of the beneficial effects of novel reading” (2010: vii; see also xiv, 20, 35, 65–6). A further problem concerning this idea of literature as a panacea is that it might lessen its appeal to the younger generation: “By advertising novels as so relevant, personally beneficial, and immediately useful to self and society, do we not render unpalatable the very product we wish to place in the hands of the young?” (2010: 64). She ironically adds: “Novels, surely, can still be sexy, time wasting, and subversive – or do they have to be vitamin-enriched bowls conveying good-for-you moral fiber?” (2010: 64). In educational settings, striking the right balance between antibiotics and food-for-thought is challenging, especially when reading is conceived of as a moral intervention.

Keen’s book is a fascinating exploration of what empathy means to someone who fully acknowledges its importance, but is still struggling with some of the repercussions on traditional reading(s) and the extent of empathy’s impact on different readers. We have seen that the right degree of identification and empathy/sympathy is a delicate balance between the author’s intentions and the readers’ actual reactions, between textual structures and the readers’ associations, between familiarity and strangeness, between critical distance and transportation etc. If readers maintain too much distance, they are less likely to become entangled; however, if they become too caught up in the story, they lose all perspective and begin to see a complex subject matter from just one perspective. Keen’s scepticism is tied to her belief that “the empirical evidence for causal links between fiction reading and the development of empathy in readers does not yet exist” (2010: 124), which makes her doubt that empathy “inevitably yields the cultural and civic good of altruism and engaged world citizenship” (2010: 145) that many teachers may have in mind. She is worried by the claims of some psychologists who compress a number of tentative speculations and circumstantial pieces of evidence into a fully-fledged theory of how reading equals social competence. To balance this one-sided view, Keen highlights several factors that may impact this skill, such as personal dispositions (cf. 2010: 72), “generic and formal choices made by authors” (2010: xii), hereditary factors (cf. 2010: 3), social upbringing (cf. 2010: 3) as well as “historical and social contexts” (2010: 81).

For Keen, this whole matter comes down to a single question: “Do empathetic people make good readers, or do good readers become empathetic people?” (2010: xv). For educational settings, there has to be hope that the latter remains a possibility, but Keen is surprisingly sceptical: “teachers in particular […] have employed narrative fiction to steer children toward greater empathy. ←173 | 174→This widespread practice raises the question of whether empathy can in fact be taught through reading” (2010: 11; see also 12). Since this is not going to happen without any effort, teachers have to take a more active role: “Reading alone (without accompanying discussion, writing, or teacherly direction) may not produce the same results as the enhanced reading that involves the subsequent discussion” (2010: 91; see also 146–7). This returns us to Sklar’s observation that empathy/sympathy may be as much a result of class work as of reading itself, which is directly compatible with aesthetic reading. According to this logic, an engagement with the text is an ongoing process and the students’ response to the literary work is equally shaped by further classroom activities. Following Sklar, a teacher’s responsibility in the area of empathy/sympathy would be less about enabling identification, which he takes for granted, but more about bridging the gulf between the self-contained world of the book and the real-world issues that are addressed.

If literature should have an impact beyond the individual readers’ forms of self-implication, the meaning of the text has to be negotiated. Keen argues that pro-social behaviour is only possible “with the guidance of a teacher who connects the dots between reactions to fiction and options for action in the real world” (2010: 146), which is exactly the point why reading has to take place as a process and a sequence of steps. It also means that teachers who preach pro-social behaviour and intercultural understanding should be prepared to take the saying ‘Actions speak louder than words’ seriously. Students may be willing to become actively involved instead of feeling complacent about having addressed the issue in theory only. The classroom is one of the few remaining public spaces of civilised discussion and negotiation, which makes the activities tied to a reading crucially important.

3.4 Embodied Cognition & Enactivism

In The Cambridge Companion to Dewey (2010) Mark Johnson offers an article entitled “Cognitive Science and Dewey’s Theory of Mind, Thought, and Language” in which he sets out to demonstrate how closely connected the two are, especially concerning embodied cognition:

Dewey also anticipates some of the most significant empirical findings of recent cognitive science research on the bodily grounding of meaning. We have seen that in Dewey’s theory of mind and thought, there is no place for ideas as quasi-entities floating around in some disembodied mental space, subject to manipulation by an allegedly pure ego. On the contrary, meaning has to come from experience, and experience is at once irreducibly bodily, biological, and cultural. (2010: 136; see also 139)

←174 | 175→

In this sense, “[a];ll thinking arises from bodily processes of organism-environment transaction, and it takes whatever value it has from its ability to enrich and transform that experience” (2010: 129). Although Dewey has always been a staple of reader-response criticism and literature in the classroom, not least through Louise M. Rosenblatt’s influential work, his teachings seem to undergo a renaissance with the rise of cognitive studies. In Marco Caracciolo’s enactivist approach to literature, for example, Dewey is a ubiquitous point of reference (cf. Caracciolo 2014: 22–3, 49, 51, 73–5, 77, 89–90).

Throughout this thesis readers have already encountered the basic premises of embodied cognition, for example in Monika Fludernik’s observation that “man’s enmeshment or engagement with his environment operates as a central constitutive feature and as a fundamental cognitive frame” (2005: 7; see also 311). Even Schank and Abelson acknowledged that schemes and scripts have their origin in actual experiences, so that the efficacy of the restaurant script, for example, largely depends on taking the role of customer again and again:

A script must be written from one particular role’s point of view. A customer sees a restaurant one way, a cook sees it another way. Scripts from many perspectives are combined to form what might be considered the ‘whole view’ of the restaurant. Such a ‘whole view’ is rarely, if ever, needed or called up in actual understanding, although it might well constitute what we may consider to be one’s ‘concept’ of a restaurant. (1977: 42)

The same logic applies to objects and tools in particular (cf. Stewart 2014: 18–21). While they “are often defined by the scripts they relate to”, objects are first encountered in an interactive manner: “We would guess that for children, the definition of the object is, apart from its physical description, identical to its place in a script. That script is defined in terms of the first experience the child had with the object” (Schank & Abelson 1977: 225). In other words: for a long time children operate with the experiential knowledge of handling an object in a specific context. Much in the same way that we learn our first language in an implicit and strongly contextualised manner, so do we learn to interact with and within specific (social) environments. Daniel Hutto explains “practical knowledge” in the following way: “I know how to tie my shoes; to ride a bike; to play table-tennis; but these abilities do not rest on a kind of propositional rule following. It follows that we cannot say, even in principle, how we achieve such feats by articulating the set of tacit rules or maxims followed since there are none” (2005: 390; see also 395). Like native speakers of languages, we become native handlers of objects. Elena Cuffari et al. explain this type of learning as follows: “Growing up in the environments-ecologies-milieus that people do, we develop sensitivities ←175 | 176→to certain acts and strategies of coping, and we incorporate the coping practices until they become constitutive of our way of being in the world” (2015: 1092).

This ties in with what John Stewart calls “a strong revival of a neo-Gibsonian ecological approach to perception” (2014: 4), by which he means James Gibson’s concept of affordances (cf. Edgar, Edgar & Pike 2014: 45). What follows is an explanation by Francisco J. Varela, Evan Thompson and Eleanor Rosch, but it has to be noted that they criticise Gibson for conceptualising affordances as characteristics of the environment instead of interactive properties:

In Gibson’s view, certain properties are found in the environment that are not found in the physical world per se. The most significant properties consist in what the environment affords for the animal, which Gibson calls affordances. Stated in precise terms, affordances consist in the opportunities for interaction that things in the environment possess relative to the sensorimotor capacities of the animal. For example, relative to certain animals, some things, such as trees, are climbable or afford climbing. Thus affordances are distinctly ecological features of the world. (1993: 203; see also Caracciolo 2014: 76)

Their argument is rather that “living beings and their environments stand in relation to each other through mutual specification or codetermination” (1993: 198). Graham Edgar, Helen Edgar and Graham Pike are more appreciative of Gibson’s contribution and see him as an early forerunner of embodiment in the field of cognitive psychology: “Gibson, rather than considering how perception operates, was much more concerned with what perception is for. That is, Gibson proposes that perception should be considered in terms of how it allows us to interact with the world we live in. […] For Gibson, moving within the environment and interacting with the environment are crucial aspects of perception” (2014: 45). In this sense, Dewey’s ‘minding’ is as much a ‘doing’, as cognition, action and emotion go hand in hand: “Not only is art itself an operation of doing and making – a poiesis expressed in the very word poetry – but esthetic perception demands, as we have seen, an organized body of activities, including the motor elements necessary for full perception” (2005: 267). Jean Piaget and Bärbel Inhelder stress that perceptual information is always stored together with interactive, sensory-motor experiences: “Neurologically, an imagined bodily movement is accompanied by the same pattern of electrical waves, whether cortical (EEG) or muscular (EMG), as the physical execution of the movement. That is, the imagining of a movement involves a kind of sketch of that movement” (2000: 68; see also 69). Even at higher levels of cognition/compression, motor memories never go away: “The connection between memories and schemes of action suggested by the preceding facts, along with the schematization of memories as such studied by F. Bartlett, makes such a reconciliation conceivable by emphasizing the importance of motor or ←176 | 177→operatory elements at all levels of memory” (2000: 83). This is how Johnson explains Dewey’s take on this matter:

Because Dewey rejects mind/body dualism, he regards the activity of thinking as just as much a matter of habits as any other form of human bodily activity. Just as when a potter employs motor skills to mould clay by means of the manual eye-hand habits she has painstakingly developed, so also the ways we think are the present result of developed and still-developing habits for working through experience. (2010: 131)

The dominant cognitive theories in linguistics are all based on embodied cognition, so it is far from a coincidence that Mark Johnson contributed to a collection of essays on Dewey’s philosophy: “The basic form of explanation is that meaning is grounded in our sensory-motor experience and that these embodied meanings are then extended, via imaginative mechanisms such as images, schemas, conceptual metaphor, metonymy, radial categories, and various forms of conceptual blending, to shape abstract thinking” (2010: 139). This is the content of the next chapter in a nutshell.

Enactivism, now, is an attempt to explain cognition “as an essential feature of living organisms” (Stewart 2014: 1) without any recourse to abstract mental models and prototypes. This “ ‘knowing how’ expressed directly in action” is conceptualised as “much more basic and much more generic than symbolic knowledge” (2014: 3; see also Hutto 2005: 389). Accordingly, enactivists explicitly position themselves in opposition to a “Computational Theory of Mind” (Stewart 2014: 1) and its two dominant strands – theory theory and simulation theory. An important foundational text is Varela, Thompson and Rosch’s The Embodied Mind: Cognitive Science and Human Experience (1993) in which the authors deplore the fact that “the computer model of the mind is a dominant aspect of the entire field” (1993: 4). On a basic level, the enactivists are worried by our western “tendency to overintellectualize our capacity to evaluate and understand” (Colombetti 2014: 147), but also by the tendency that the (natural) sciences have lost all contact to what human life is about. They even demand a return to what they refer to as “common sense” (1993: 148; see also 149), which is human sense-making in contrast to computational models:

The term hermeneutics originally referred to the discipline of interpreting ancient texts, but it has been extended to denote the entire phenomenon of interpretation, understood as the enactment or bringing forth of meaning from a background of understanding. In general, Continental philosophers, even when they explicitly contest many of the assumptions underlying hermeneutics, have continued to produce detailed discussions that show how knowledge depends on being in a world that is inseparable from our bodies, our language, and our social history – in short, from our embodiment. (1993: 149)

←177 | 178→

It has to be stated that – contrary to a later radicalisation of the concept (cf. Hutto 2005) – it was originally intended to form a counterweight to the western tradition by “complementing cognitive science with a pragmatic, mindful, open-ended approach to human experience, such as we find in the mindfulness/awareness tradition” (Varela et al. 1993: 53) of Buddhism. We have already encountered ‘minding’ and ‘noticing’ as relevant to learning, especially in the context of defamiliarisation and foregrounding. However, Varela et al. deny System 1 operations altogether and declare that cognition is always conscious (cf. 1993: 49–51).

Another important aspect of enactivism is that it attempts to explain basic, everyday interactions as encounters between an organism as an adaptive system and an environment: “our experience is always changing and, furthermore, is always dependent on a particular situation. To be human, indeed to be living, is always to be in a situation, a context, a world. We have no experience of anything that is permanent and independent of these situations” (1993: 59). Thus, they “propose as a name the term enactive to emphasize the growing conviction that cognition is not the representation of a pregiven world by a pregiven mind but is rather the enactment of a world and a mind on the basis of a history of the variety of actions that a being in the world performs” (1993: 9). At this point it becomes obvious that enactivism attempts to operate largely on the basis of episodic memories: “all mental states (perception, memory, etc.) are of or about something” (1993: 15), which play a central role when the same situation arises again. Most importantly of all, cognitive phenomena are all tied to bodily experiences, as there is no perception, cognition or emotion that is not mediated through the body (cf. 1993: xvi, 65, 173). They reject the possibility “that information exists ready-made in the world and that it is extracted by a cognitive system” (1993: 140), as all information is bound to contexts and specific perspectives. Hanne De Jaegher and Ezequiel Di Paolo explain that “organisms cast a web of significance on their world” and “actively participate in the generation of meaning in what matters to them; they enact a world. Sense-making is a relational and affect-laden process grounded in biological organization” (2007: 488). This sounds suspiciously like a “radically constructivist” approach (Stewart 2014: 27), which is indeed what many enactivists are drawn to.

What is the added value of discussing enactivism in the context of teaching English? First of all, there is a direct link between the foundational texts of literary reading in the classroom, especially Dewey’s Art as Experience and Rosenblatt’s transactional theory, and the concept of grounding cognition in subjective experience without denying the influence of cultural practices. In this sense, we find an interesting revival, re-evaluation and recontextualisation of Dewey’s ideas in ←178 | 179→modern philosophy. Enactivists also subscribe to what Lev Vygotsky calls the zone of proximal development (cf. Vygotsky 1966: 103), as organisms can only deal with the next natural step in their development and instruction is largely pointless if it does not respect the stage the learner is in (cf. Stewart 2014: 9; Di Paolo et al. 2014: 44). Contrary to Vygotsky, who heavily relies on instruction and the transmission of knowledge and skills, enactivists understand social interaction as a dialogue, what De Jaegher and Di Paolo call ‘participatory sense-making’ (cf. De Jaegher & Di Paolo 2007), during which both individuals contribute to a shared or blended space that produces emergent meaning. In any conversation we have to seek the common ground and find a common solution: “We don’t experience the other-in-interaction as totally obscure and inaccessible, nor as fully transparent (like an object fully constituted by my sense-making activity), but as something else: a protean pattern with knowable and unknowable surfaces and angles of familiarity that shapeshift as the interaction unfolds” (2007: 504). This means that misunderstandings are a natural part of communication as interlocutors can never completely know the other and predict everything that is going to happen. Cuffari et al. embrace misunderstandings as a key component of learning: “We see misunderstanding as a productive engine for renegotiating meaning and for going further in meaning sharing” (2015: 1120). Script-based interactions are an attempt to reduce the complexity of encounters by having humans take on predetermined roles, such as customer and waiter, but classroom interactions eventually have to allow for ‘participatory sense-making’. This may seem time-consuming and far less efficient than the transmission model, but the exact opposite is the case: in participatory sense-making both sides can benefit from the encounter. Students gain access to and comprehend a new situation or narrative only in terms of their previous experiences, so barring that route complicates or even hinders their reading process. This is why the seven stages of reading presented in part 2 contain as many interactive phases as possible.

Enactivism is a ‘learning by doing’ approach based on optimal challenges, which means that the enactivists endorse “pretend play” (Di Paolo et al. 2014: 76) in the social sphere. Although the situation is based on make-believe, it allows for the “construction of new environmentally and bodily mediated meaning” (Di Paolo et al. 2014: 76). More generally speaking, learning is based on active engagement and exploration:

Traditional distinctions between action and perception arise only as the specialisation of phases in an act of sense-making. Several examples that illustrate this point have been discussed in the enaction literature, but perhaps the simplest and clearest one is that of perceiving the softness of a sponge […]. The softness of a sponge is not to be found ‘in it’ but in how it responds to the active probing and squeezing of our appropriate ←179 | 180→bodily movements (e.g., with the fingers or the palms of the hand). It is the outcome of a particular kind of encounter between a ‘questioning’ agent with a particular body (sponges are solid ground for ants) and a ‘responding’ segment of the world. The confluence of lawful co-variations in this dialogue stabilises the cogniser’s sense-making into an object. Movements are at the centre of mental activity: a sense-making agent’s movements – which include utterances – are the tools of her cognition. (De Jaegher & Di Paolo 2007: 489)

Since utterances are conceptualised as (dialogic) movements in this theory, Cuffari et al. expand on this notion in a later article, which looks at “languaging as adaptive social sense-making” (2015: 1089). In this sense, “languaging is a special form of social agency, through which linguistic sense-makers negotiate between interactive and self-directed metaregulation of moment-to-moment living and cognizing” (2015: 1113). Thus, enactivists tie language-learning to sense-making in successively more involved encounters, so that language skills are built from the ground up as any other ability (cf. 2015: 1089).

Languaging as the reflexive and reflective negotiating with one’s self as with another is the ‘seed’ ability out of which abstraction, imagination, and reasoning grow, as one’s sense-making powers incorporate the moves, perspectives and expectations of others, and the horizons of significance in which they are embedded. (2015: 1110)

One important consequence of this approach is that very young children can be competent language users and communicators, although – from a cognitivist perspective – they do not have established the mental models yet that supposedly guide all cognition (cf. 2015: 1091). This is how Cuffari et al. summarise this idea:

an enactive approach to languaging will explain it as a kind of sensemaking, i.e. a way that human organisms monitor, evaluate, regulate and organize their existence. An enactive approach will relate languaging to self-produced identities and to the regulation of coupling with environmental domains that support those identities. Languaging will be understood as a way of living. It is this way of living, rather than a theory of mind, that speaking children share with other human interlocutors. (2015: 1092)

Di Paolo et al. argue that organisms “participate in the generation of meaning through their bodies and action often engaging in transformational and not merely informational interactions” (2014: 39). In other words: students do not just pass on and absorb information, they are more holistically involved and change as people – rather than as hard drives – through their interactions with peers and teachers. This is especially true of their bodies, which also have to be involved – rather than exclusively focusing on their brains (cf. Di Paolo et al. 2014: 42). Colombetti observes that “living systems necessarily establish a ←180 | 181→point of view, and moreover a concerned point of view that generates meaning” (2014: 148), so that “meaning is always relational” (2014: 148). That is why all meaning-making has to start with the students’ individual responses. In a more radical variant of enactivism, there is no objective meaning anyway, as the word ‘meaningful’ itself suggests that it is always relational. A group of people can agree on a particular meaning, of course, but then it is still meaningful to them. Enactivism was never intended to be an approach to reading, as the whole point of this philosophy is to explain how organisms – and not even humans at that – manage to adapt to and interact with their natural environments in the most basic ways. However, nothing spurs human imagination more than a lost cause.

Mario Caracciolo’s enactivist approach to reading literature is far from radical as, early on, he is “ready to concede that language-based cognition does involve mental representations. This is why a psychologically realistic theory of narrative cannot do without mental representations” (2014: 35; see also 9). Following this logic, he is more interested in a reorientation of rather than an abrupt break with cognitive literary studies (cf. 2014: 3). His compromise takes the following form: he borrows the metaphor of entanglement from Wolfgang Iser (cf. 2014: 36) to explain how self-implication takes place by “having one’s past experiences entangled in the process of reading” (2014: 34). Instead of mental models, Caracciolo relies on episodic memory and experientiality in Fludernik’s sense (cf. 2014: 3, 9, 47–9). The other component is Oatley’s ‘simulation theory’ (cf. 2014: 32, 94, 130–1), which blends readers’ own experiences with those of characters. Here is Caracciolo’s own outline of his approach:

Two psychological mechanisms play a role in this process: the first is the triggering of memories of past experiences – experiential traces, as they are known in the psycholinguistic literature. The second is mental simulation, which allows readers to put together past experiential traces in novel ways, therefore sustaining their first-person involvement with both fictional characters and the spatial dimension of storyworlds. All in all, there is a two-way movement between the background and narrative: like experiential machines, stories need experiential input, but also produce some output, since they can bring about a restructuring of each reader’s experiential background by generating new ‘story-driven’ experiences. (2014: 5)

This means that narratives have the power to tap into and activate readers’ episodic memories, which makes the experience more personal and emotionally satisfying (cf. 2014: 34, 36, 50). Yet, this is also important for another reason, as “stories cannot represent the characters’ experiences, but only events and actions whose experiential dimension is supplied by readers through their own familiarity with experience” (2014: 23). He is more explicit about this point later in his book, when he applies reader-response criticism’s concept of the text as a ←181 | 182→blueprint to the (re)construction of emotional states: “There is no stand-in for the character’s experience here, but only linguistic-expressive-pointers that cue readers into attributing an experience to a fictional being because of their own familiarity with bodily experience” (2014: 104).

Caracciolo attempts to model a progression from “consciousness-attribution” to “consciousness-enactment” (2014: 25, 41, 110), which corresponds to a step from Theory of Mind in Palmer’s sense to Oatley’s simulation, with a much stronger focus on experientiality instead of mental models. This becomes obvious in the following statement: “consciousness-attributions are not based on reasoning, but on the identification of expressions of consciousness” (2014: 117). We learn how to attribute conscious states to characters based on real-life encounters:

… primary intersubjectivity enables us to grasp another person’s intentions in an embodied, online way through face-to-face interaction, without any need to infer them via a theory or by running mental simulations. This capability is closer to perception than to higher-order cognitive processes: […] Primary intersubjectivity points to our ability to understand other people’s minds as directly accessible through their bodily behavior (for example, facial expressions and gestures). (2014: 142)

In other words, these are System 1 operations based on general experience. For consciousness-enactment, Caracciolo combines self-projection, role-taking, perspective-taking, self-implication, self-modification, sympathy and transportation into his own variety of empathy, which becomes obvious in the following passage:

… consciousness-enactment always involves an element of consciousness-attribution: consciousness-attribution brings in its wake a third-person stance toward a character, but in enacting a character’s experience readers imaginatively ‘try it on’ without completely giving up their third-person perspective (since the character always remains another subject). (2014: 49; see also 118)

He clarifies this point later in the book: “consciousness-enactment tends to build over time in the presence of the appropriate textual cues: it is a cumulative process that can spread across different ‘threads’ of our experiential engagement with narrative […] The upshot is that consciousness-enactment comes in degrees” (2014: 124). However, it also goes back and forth between more and less engagement, as readers’ perspectives constantly change as the narrative progresses. This sounds a lot like Iser’s coordination of perspectives and partial identification with a character’s point of view: “in consciousness-enactment, the recipient’s story-driven experience is made to overlap, if only partially, with the experience attributed to a character. Consciousness-enactment, therefore, creates a tension ←182 | 183→between the first-person undergoing of an experience and its second- or third-person attribution to a fictional being” (2014: 110; see also 122). Because of the wide spectrum of reader responses that have been subsumed under the umbrella term of Theory or Mind, such accounts of ‘consciousness-enactment’ tend to be rather confusing, as they involve anything from narratological analysis via simulation, perspective-taking and self-implication to enactivism.

When Caracciolo addresses what he considers to be the extreme point of the spectrum – perspective-taking – he follows an argument similar to Kuiken and Miall’s concept of self-implication and self-modifying feelings: “stories can have a feedback effect on interpreters’ experiential background at this level by inviting them to revise – in a more or less self-conscious way – their views and outlook on the world. Narrative can affect people not only imaginatively and emotionally, but also cognitively and culturally” (2014: 67; see also 141–2). This becomes even more obvious in the following statement, where appraisals are foregrounded as the major type of engagement with narratives: “Experience is not simply about reproducing the world as if from a detached, observer position. It is an online, engaged, embodied evaluation of ‘what is at stake’ in the world for creatures like us” (2014: 52; see also 64). Damasio’s argument about the inseparability of cognition and emotions is very evident in this context: “readers’ engagement with narrative texts is shot through with sensory images, emotions, evaluations – the stuff our experiences are made of, and that cannot be adequately accounted for within a computational model of the mind” (2014: 46). This experientiality that readers share with authors becomes the background against which the fictional characters’ lives are judged. Yet, to enable readers to get ‘a feel’ for the characters’ experiences – which they do not have as assemblages of signs – writers use specific expressive strategies to guide a process of increasing entanglement, which can take slightly different directions depending on the experiences individual readers bring to the text (cf. 2014: 49).

One way to do that is to rely on conceptual metaphors (cf. 2014: 21), which is the topic of the next chapter. This means that widely shared human experiences have found their way into language in the form of image schemas, basic metaphors and all the other figures of speech that build on these foundations. Language can trigger physical responses to imaginary scenarios as long as it manages to tap into our personal experiences: “we just have a gut reaction to the spider as soon as we imagine it crawling up our arm. What is remarkable about this phenomenon is the way mental and semiotic representations can draw on a level of our engagement with the world that is prelinguistic and non-representational” (2014: 38). In this sense language mainly works metonymically and not representationally: it is not so much the linguistic ‘depiction’ of a large, hairy spider ←183 | 184→crawling up one’s arm, but language’s power to evoke – or provoke – a response, based on real-life experiences.

An important question in this context is the problem of how “the phenomenal qualities (or ‘qualia’) of bodily-perceptual experiences” (2014: 24) can be expressed through language, to which Caracciolo gives the following answer: “metaphors and similes are one of the expressive devices through which stories invite readers to enact the felt qualities of specific experiences in a way that […] can have a deep impact on their background” (2014: 109). It becomes more interesting when he elaborates on how this takes place:

Metaphorical language adds a layer of complexity to this indexing process [language cueing experiences]: in order to convey the qualia of a bodily-perceptual experience, it points to another experience (usually, but not necessarily, at the same level of the background), asking the reader to conflate the two. The qualia conveyed to the reader correspond to what Fauconnier and Turner (2002) would call the “emergent meaning” of the metaphorical mapping – the value added that results from blending the two experiences. (2014: 108)

This is remarkably similar to Miall and Kuiken’s approach presented above, who speak of “a form of enactive reading that implicitly blends the fictional world with what readers know, believe, or feel about their own lives. […] In these cases, the reader is, we suggest, confronting personal feelings and recontextualizing them in the light of the fresh feelings evoked during reading” (2002: 238). The feelings attributed to or projected onto the character – what Miall and Kuiken call ‘remembered emotions’ and Caracciolo ‘experiential traces’ (cf. 2014: 46) – have to come from the reader, which may vary in their specificity. The unique longing for a cigarette may invite heavy smokers to instantly bond with the character, but everyone else is also familiar with cravings or, even more broadly, not getting what one wants, which may be sufficient for a basic understanding of the situation. Thus, we flesh out the characters with standard schematic knowledge that we automatically apply without much thought. Yet, art being art, the specific configuration of circumstances and the disposition of the character may invite what Miall and Kuiken call self-implication. The easier it is to relate to a character or situation, the more likely it is that the narrative can tap into our personal memories, which adds more flavour to the experiences we bring to the reading: “it is important to keep in mind that readers do not just attribute mental states to fictional characters – they attribute mental states with a qualitative aspect or with qualia in the broad sense” (Caracciolo 2014: 112).

This returns us to reader-response criticism’s blueprints and musical scores: literary texts use cues and foregrounding to trigger reactions, associations, ←184 | 185→appraisals, blends etc. but they do not represent reality in and of themselves or predominantly trigger world building. However, foregrounding occurs in an organised and orchestrated fashion, so that the gaps in narratives are highly productive. Therefore, I disagree with Caracciolo’s suggestion that the gaps are qualitatively the same in both instances: “My argument is that, despite common assumptions to the contrary, the perceived world is as sketchy and ‘gappy’ as the mental imagery generated during the reading of narrative texts. This brings grist to the mill of my account of the structural resemblance between real experiences (perception) and story-driven experiences (imaginings)” (2014: 24). Art is much more ‘expressive’ – to use Caracciolo’s own term, teleological and arranged. The power of art is to draw us in and invite self-implication, whereas real life is largely marked by the comfortable boredom of daily routines. There is no noise in the information – especially not in the most condensed verbal arts like poetry – where every single word is carefully chosen; but the same principle applies to all forms of artistic expression. In real life, people do not emote enough and we are often challenged to interpret the complexity of other humans’ emotional states without sufficient clues. In art, facial expressions, gestures, body postures etc. are all carefully selected to be legible – or not, which is in itself meaningful, as everything we read is based on deliberation. When Caracciolo claims that narratives can ‘express’ an experience through such signs, which limits the range of potential responses, this is only possible because of art:

… insofar as stories draw on, and have an effect on, the experiential background of recipients, they can also express an experience by having recipients respond to the represented events and existents in certain ways. […] fictional consciousnesses are (just like the consciousnesses of real people) attributed on the basis of external signs, such as gestures and psychological language, which we take as expressive of an experience. Consciousness-attribution is thus readers’ most basic strategy for dealing with fictional consciousnesses. (2014: 114)

This does not mean that we do not read an angry face in real life and in fiction in largely the same way – especially in visual narrative media such as film or comics – but if the point of readers’ engagement with texts is the reconstruction of fictional consciousnesses based on these signs, there is presumably more guidance in fiction. This is also what Caracciolo ultimately agrees to:

I would like to leave the door open for the idea that readers’ access to other consciousnesses can be more direct when dealing with fictional characters than with real people. My proposition is that our engagement with fictional consciousnesses differs in degree, but not in kind, from our engagement with real ones, in particular when it comes to consciousness-enactment: we can enact the experience of another person in ←185 | 186→real life too – that is to say, empathize with him or her – but we do not do it as often (and as intensely) as in reading texts … (2014: 113; see also 129)

The reason is that “[s];ympathy and empathy for characters (together with the related stances […] ‘consciousness-attribution’ and ‘consciousness-enactment’) can be modulated through sophisticated expressive techniques” (2014: 66; see also 143). These operate within a number of large frameworks, such as sharing experiences with the writer through joint attention, “evolutionary, cultural and personal dispositions” (2014: 42) or genre knowledge. On the discourse level, Caracciolo argues, we encounter authors’ or narrators’ explicit evaluations and attitudes towards characters and situations, stylistic foregrounding, narrative structure, psycho-narration, focalisation, and characters’ points of view (cf. 2014: 42–4, 104, 126–9), which all offer a carefully calibrated guide as to how readers are supposed to read and identify with characters. These

… expressive devices at the micro-level of analysis […] occur frequently in narrative texts. However, it is only the accumulation of these devices (i.e., when a story gives sustained attention to a character’s experience) that primes consciousness-enactment: a coordinated set of low-level triggers constitutes a higher-level trigger, inviting interpreters to enact a character’s consciousness throughout a larger textual unit. This process culminates in what I call “consciousness-focused narration”. (2014: 125; see also 41, 128)

The accumulation of cues, textual evidence or what Smith would call the in-built redundancy of texts (cf. 2004: 63–5), steers the readers in particular directions. Concerning internal focalisation, perspective-taking can never be complete, as “a consciousness is not a place from which we experience the world – it is, first and foremost, the medium through which we experience it [; …] a consciousness is not an object-like entity; it cannot be reduced to a spatio-temporal position from which we imagine some events and existents” (2014: 118). Again, this is tied to the idea that literature does not mimetically represent experiences as quasi-objects in literary texts, which readers take in as a piece of information: “characters’ consciousness and experiences cannot be represented as such by narrative texts; what we commonly call the ‘representation of an experience’ is the representation of an event in which a person (e.g., a fictional character) undergoes an experience” (2014: 30; see also 31). This is not to say that characters’ emotions cannot be accessed almost instantaneously, but this recognition usually relies on the somewhat circumlocutionary use of metaphor, metonymy and situational context rather than direct access. Readers have to share their own experiences and episodic memories with the characters to grant them an emotional depth that cannot objectively exist in the text.

←186 | 187→

What hampers Caracciolo’s otherwise fascinating attempt to make enactivism compatible with literary studies is a twofold problem with a single root: despite the implicit claim that enactivism is a new paradigm that revolutionises the way we conceptualise cognition and reading in particular, there is a strong reliance on classical narratology and conventional approaches within cognitive literary studies. From the outset, Caracciolo acknowledges that he embraces Theory of Mind, especially simulation theory, which is based on mental representations and schema theory. When he addresses empathy and perspective-taking, traditional conceptualisations of focalisation based on textual clues become very prominent. The other problem concerns the case studies: they are rather short and not all that convincing. If we take Caracciolo’s reading of Ian McEwan’s On Chesil Beach (cf. 2014: 144–51) as an example, there is some new terminology, but very little else that cannot be explained through traditional means or Alan Palmer’s social minds.

Despite these shortcomings, Caracciolo’s book is still an important effort to build literary reading from the ground up and expand narratology beyond its narrow confines. He highlights the continuities between John Dewey, Wolfgang Iser, cognitive literary studies and enactivism in particular, which opens up a diachronic perspective and reveals a sensibility to continuities that is frequently missing from other publications in the same vein, such as Alan Palmer’s Fictional Minds (2004) or Barbara Dancygier’s The Language of Stories (2012). Episodic memories, feelings and the experientiality of everyday life become readers’ key resources in making sense of literary texts and the entanglements of characters in particular social configurations. Thus, he reinforces John Dewey’s, Frank Smith’s and Monika Fludernik’s insistence on the continuities between real life and art. To build on Carraciolo’s groundwork and to extend this approach to visual narrative media, we have to turn our attention to linguistics first.

3.5 Conceptual Metaphors & Blending

3.5.1 Basic Principles

Cognitive linguistics is essential to an experiential approach to literature and comics in particular, for which this chapter provides the theoretical foundation. The field’s “Cognitive Commitment” is meant to ensure that the “principles of linguistic structure should reflect what is known about human cognition from other disciplines, particularly the other cognitive sciences” (Evans & Green 2006: 40). Accordingly, one can detect numerous parallels to the theories presented thus far.

←187 | 188→

The first central tenet of cognitive linguistics is a fierce opposition to “language as a formal or computational system” (Evans & Green 2006: 44), which Vyvyan Evans and Melanie Green also call the “empiricist view” (2006: 44) or the “objectivist approach” (2006: 47). In this context, Lakoff and Johnson observe that the “set-theoretical concept of a category does not accord with the way people categorize things and experiences. For human beings, categorization is primarily a means of comprehending the world, and as such it must serve that purpose in a sufficiently flexible way” (2003: 122). This leads them to the conclusion that “at least some of the properties that characterize our concept of an object are interactional. In addition, the properties do not merely form a set but rather a structured gestalt, with dimensions that emerge naturally from our experience” (2003: 122). This is quite a departure for a linguistic approach, in that the meaning of words is neither in the text, nor in the brain, but emerges naturally through an interactive process. In enactivism, this is the foundation of all cognition. The prototypes we operate with – e.g. the words we use – are not sets of characteristics, but tentative, experiential gestalten (cf. 2003: 77–86, 210), very much in Iser’s sense, or ‘blends’ in conceptual integration theory. They are never ‘pure’, objective and shiny, but used, rough round the edges, somewhat fuzzy and ‘tainted’ by experiences and emotions. They are organised according to family resemblances (cf. 2003: 71, 122–3; Kövecses 2010: 109), rather than marked off by distinct sets of attributes. Experiential gestalten as “ways of organizing experiences into structured wholes” (2003: 81) allow for a holistic approach, which is essential, as “we need to classify our experiences in order to comprehend, so that we will know what to do” (2003: 83). The brain is naturally prone to ‘organic’ networking and establishing links between seemingly incompatible things, but less so to rational, ‘mechanical’ analysis in the sense of taking apart and looking at the functional value of items (System 2).

Since the acquisition and use of language have to be explained on the basis of general cognitive principles rather than a specific language module in the brain (cf. Evans & Green 2006: 41), George Lakoff affirms that “language is secondary” (2007: 273), meaning that we have to look at pre-linguistic, subconscious phenomena first, or in other words: at basic human experiences as the source of linguistic expressions. In the second chapter of their massive Cognitive Linguistics: An Introduction, entitled “The nature of cognitive linguistics: assumptions and commitments”, Evans and Green start with what they consider to be the foundational thesis of the discipline: “embodied cognition” (2006: 27). This mirrors enactivism’s basic premise that “the concepts we have access to and the nature of the ‘reality’ we think and talk about are a function of our embodiment: we can only talk about what we can perceive ←188 | 189→and conceive, and the things that we can perceive and conceive derive from embodied experience. From this point of view, the human mind must bear the imprint of embodied experience” (2006: 46; see also 44; Lakoff 1990, xiv-xv).

Despite the fact that humans develop a fairly sophisticated and wide-reaching network of interrelated conceptual metaphors, the basic building blocks are primary metaphors that are based on direct bodily experiences (cf. Kövecses 2010: 7) and “arise spontaneously and automatically without our being aware of them” (Lakoff & Johnson 2003: 256). Here is how Lakoff and Johnson explain this ‘subindividual level’ (cf. Kövecses 2010: 309–11) that forms the basis of all cognition before culture and language begin to influence and modify a person’s conceptual system:

Since the mechanism of metaphor is largely unconscious, we will think and speak metaphorically, whether we know it or not. Further, since our brains are embodied, our metaphors will reflect our commonplace experiences in the world. Inevitably, many primary metaphors are universal because everybody has basically the same kinds of bodies and brains and lives in basically the same kinds of environments, so far as the features relevant to metaphor are concerned. (2003: 257; see also 119, 257, 259; Kövecses 2010: xi-xii; 195–213)

This includes basic spatial orientation (cf. 2003: 56–7; see also Kövecses 2010: 88) and the handling of objects, which have interactional properties (also called affordances) that invite or allow for certain types of engagements (cf. 2003: 120, 214–5). Thus, embodied cognition establishes the basic parameters for further experiences and linguistic expressions: “our experiences with physical objects (especially our own bodies) provide the basis for an extraordinarily wide variety of ontological metaphors” (2003: 25) and “are characterizable as multidimensional gestalts whose dimensions emerge naturally from our experience in the world” (2003: 121–2; see also 123, 162). An important component of this process of perpetual interaction is instant feedback: “our conceptual system emerges from our constant successful functioning in our physical and cultural environment” (2003: 180).

Although conceptual metaphor theory seems to be similar to enactivism in this respect, there are three significant differences. Despite the shared tenet that primary experiences are generated through direct interactions between organism and environment, cognitive linguistics operates with mental frames. Secondly, cognitive metaphor theory stresses that new insights can be gained by metaphorical thought – or blending – independent of immediate contexts, which includes creative/artistic work. Lakoff and Johnson argue that “meaning is negotiated”, which becomes the prerequisite for “an interactionally based and ←189 | 190→creative understanding” (2003: 231). Thirdly, they fully acknowledge automated System 1 operations, which are anathema to enactivists: “what is deeply entrenched, hardly noticed, and thus effortlessly used is most active in our thought” (Kövecses 2010: xi; see also 34).

Our most basic interactions with the environment lead to the development of what cognitive linguists call ‘image schemas’ (cf. Evans & Green 2006: 46–7). These are “abstract concepts consisting of patterns emerging from repeated instances of embodied experience” (Evans & Green 2006: 179; see also 176). A simple example is the container image schema (cf. Evans & Green 2006: 158). We begin to understand our own bodies as bounded objects or containers with things going in and out, but also our environment presents numerous examples of things contained within or passing into and out of other things. object is another essential image schema, but there are dozens, such as path, near-far, up-down, part-whole, full-empty etc. Their experiential basis cannot be restricted to visuals only, but includes all sensory perceptions (cf. Evans & Green 2006: 179; see also Oakley 2007: 216). In our conceptual system image schemas are “the foundations” (Evans & Green 2006: 180), as we use these basic, often spatial orientations to make sense of more abstract concepts via “metaphorical projection” (Evans and & Green 2006: 158). Not only is the body the source of our experiences, but also of our metaphors: “We are physical beings, bounded and set off from the rest of the world by the surface of our skins, and we experience the rest of the world as outside us. Each of us is a container, with a bounding surface and an in-out orientation” (Lakoff & Johnson 2003: 29). Therefore, Lakoff and Johnson claim that we tend to conceive of everything as containers, starting with our own minds, which we can feed with new ideas.

The container image schema leads to an understanding of our heads/brains as containers and of ideas as objects that enter the brain and come out again through our mouths in the form of language – ‘containing’ the idea in a sentence. In Charles Dickens’s Hard Times Thomas Gradgrind, the schoolmaster, strongly believes in the idea of feeding facts to his students, the “little vessels”, who are supposed to get “ready to have imperial gallons of facts poured into them until they were full to the brim” (Dickens 1995: 9). This widespread misconception of how learning and meaning-making take place provided the impetus for one of the most important publications in cognitive linguistics, George Lakoff and Mark Johnson’s Metaphors We Live By.

In their preface they explain that the main motivation to write this book was the realisation “that ‘meaning’ in these traditions [Western philosophy and linguistics] has very little to do with what people find meaningful in their lives” (2003: ix), which is closely tied to the necessity to “understand their experiences” ←190 | 191→(2003: 116). So, naturally, “meaning is always meaning to someone. There is no such thing as a meaning of a sentence in itself, independent of any people” (2003: 184; see also 197, 205). It is also “not cut and dried; it is a matter of imagination and a matter of constructing coherence” (2003: 227) through inference (cf. 2003: 244). From these few statements alone it should become obvious that reader-response criticism and cognitive linguistics share a lot of common ground. When Evans and Green explain that “meaning arises from a dynamic process of meaning construction, which we call conceptualisation” (2006: 363), this argument is easily compatible with Louise M. Rosenblatt’s transactional theory (cf. 1994). Since Evans and Green do not believe in the intrinsic meaning of words, they always consider the entire contextual frame: “conceptualisation is guided by discourse context, which forms an integral part of the meaning construction process. According to this view, meaning construction is localised and situated, which entails that pragmatic (context-dependent) information and knowledge inform and guide the meaning construction process” (2006: 367). Looking at the role of language, we find an uncanny similarity to Dewey’s “musical score” (2005: 113), Rosenblatt’s “blueprint” (1994: 86, 88) or Iser’s “instructions for the production of the signified” (1980: 65):

… cognitive semanticists argue that words are prompts for meaning construction rather than ‘containers’ that carry meaning. Furthermore, according to this view, language actually represents highly underspecified and impoverished prompts relative to the richness of conceptual structure that is encoded in semantic structure: these prompts serve as ‘instructions’ for conceptual processes that result in meaning construction. In other words, cognitive linguists argue that meaning construction is primarily conceptual rather than linguistic in nature. (2006: 214; see also 366, 368, 371; Coulson & Oakley 2000: 176)

Thus, the power of words stems from their artful arrangement and their ability to activate the mental capacities of readers or listeners. The process of meaning-making, which Iser calls gestalt-forming, does not primarily involve the creation or update of mental models, but the conceptual integration of several input spaces. With cognitive linguistics – as with reader-response criticism – the focus shifts from the containers, such as words, sentences, texts, mental models etc., to their interrelations. These links between conceptual spaces – what cognitive linguists call ‘mappings’ – become the central concern.

There are three basic ways of relating conceptual entities to each other: metonymy, metaphor and conceptual integration, which has already been introduced as blending. In the first case, both of the concepts belong to the same mental space and one activates and leads to the other. According to this logic, a cross may refer to Christianity, a logo to a brand, a stethoscope or ←191 | 192→white coat to a doctor, a passport photograph to a person and so on. In each of these cases, we have a relatively close and conventionalised relationship that quickly calls the other concept to mind. Metaphor, however, requires two different, often very distant frames. Mapping – or “conceptual projection” (Evans & Green 2006: 286) – takes place when we use a more concrete, closer-to-daily-life frame to make sense of a more abstract idea or concept, which means that “conceptual metaphors are mostly unidirectional” (Kövecses 2010: 27; see also Evans & Green 2006: 296–7). Since the two domains are not directly related, “mappings are asymmetric and partial” (Lakoff 2007: 309), so that only certain structures are mapped – based on their shared image schemas (cf. 2007: 296). As its name reveals, conceptual metaphor theory is more concerned with the second type. To acknowledge that fact, metaphors come first, before we return to metonymy and finally look at blending in greater detail.

3.5.2 Metaphors

The previous section ended with a tentative definition of metaphor, which now requires an example to illustrate the type of conceptual projection that takes place. time is money uses the more familiar experience of managing a resource to structure an understanding of time according to the same logic. However, although we can ‘save’ time, we cannot take it to the bank and get dividends at the end of the year. That is why “cross-domain mappings” (Evans & Green 2006: 286) only shed light on certain aspects of the target domain and keep others in the dark (cf. Lakoff & Johnson 2003: 10). This ‘metaphorical highlighting’ (cf. Kövecses 2010: 91), which “necessarily goes together with hiding” (Kövecses 2010: 92), is the reason “why a single target concept is understood via several source concepts: one source just cannot do the job because our concepts have a number of distinct aspects to them and the metaphors address these distinct aspects” (Kövecses 2010: 135). In this sense, source domains work like prisms and create a certain perspective, such as introducing capitalist notions to human existence in the form of time management. This aspectual quality of metaphors is very similar to literature in general, where – according to Iser – perspectival structures also offer particular angles that readers have to navigate and negotiate. The power of metaphors lies in the tension between surprising revelations through the joining of vastly different frames and the jarring incongruities that such a creative mapping produces. Thus, metaphor “is one of our most important tools for trying to comprehend partially what cannot be comprehended totally: our feelings, aesthetic experiences, moral practices, and spiritual awareness” (Lakoff & Johnson 2003: 193). Accordingly, metaphorical ←192 | 193→thinking constitutes a process of approximation, a means of getting to grips with elusive, ephemeral qualities that are difficult to hold on to or to make sense of. It is creative, poetic and far from perfect. That is why we need several conceptual metaphors to grasp a single abstract concept – depending on the context and a particular person’s experiences and approach. However, there are countless metaphors that have become so conventionalised that we take them for granted instead of questioning what they hide.

The idea that mental spaces, models, domains and frames are containers, bounded entities with fixed elements in them, is just a convenient way of talking about complex cognitive processes in terms of the familiar. There is no doubt about the network structure of long-term memory, so the question which box exactly contains one’s understanding of what a hammer is, remains unanswerable, as there are no containers in the first place. Cognitive linguistics comes with its own set of labels for these conceptual spaces. In cognitive metaphor theory, for example, the preferred term seems to be ‘domain’: a “conceptual metaphor consists of two conceptual domains, in which one domain is understood in terms of another. A conceptual domain is any coherent organization of experience” (Kövecses 2010: 4). In cognitive semantics, Charles Fillmore has established the label ‘frame’, which puts more emphasis on the relations than the containers themselves:

By the term ‘frame’ I have in mind 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; when one of the things in such a structure is introduced into a text, or into a conversation, all of the others are automatically made available. I intend the word ‘frame’ as used here to be a general cover term for the set of concepts variously known, in the literature on natural language understanding, as ‘schema’, ‘script’, ‘scenario’, ‘ideational scaffolding’, ‘cognitive model’ or ‘folk theory’. (2007: 238; see also Evans & Green 2006: 166, 211, 222)

For Fillmore, all the computational terms are insufficient as they disregard vital links between concepts and their contexts. The meaning of a thing is not an entity, but a network, so there is no possibility to speak about meaning in cognitive semantics without looking at the larger picture. Meaning is not of the kind recorded in dictionaries – isolated words arranged in an organised but otherwise arbitrary fashion – but of encyclopaedias or words in context (cf. Evans & Green 2006: 206–47). Lexical concepts maintain a figure-and-ground relationship to the background frame (cf. Evans & Green 2006: 222), so that their meaning only comes into sharper focus against the background and both depend on each other for their relevance (cf. Fillmore 2007: 243). They maintain metonymic relationships and evoke each other: “in the process of using a language, a ←193 | 194→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” (Fillmore 2007: 246; see also 249; Evans & Green 2006: 160–2). This becomes especially relevant in the context of genres: “knowing that a text is, say, an obituary, a proposal of marriage, a business contract, or a folktale, provides knowledge about how to interpret particular passages in it, how to expect the text to develop, and how to know when it is finished” (2007: 243) This suggests that there is a metonymic relationship between genre markers and the whole structure, as in the case of ‘film noir’, ‘private eye’ and ‘femme fatale’. However, Fillmore’s concern is predominantly with more standard situations. A ‘vegetarian’ is not just a person who does not eat meat, but evokes a whole cultural context of dietary customs and expectations (cf. 2007: 245). To Fillmore, “words represent categorizations of experience” (2007: 238) and are schematic in nature, “a skeletal representation of meaning abstracted from recurrent experience of language use” (Evans & Green 2006: 216). The actual “meaning is a process rather than a discrete ‘thing’ that can be ‘packaged’ by language” (Evans & Green 2006: 162).

This is why the codified or lexicalised ‘meaning’ of words is idealised or prototypical rather than concrete and immediately accessible. In Women, Fire, and Dangerous Things Lakoff uses the term ‘idealized cognitive model’ (ICM) for prototypical or stereotypical schemas, such as ‘mother’ or ‘marriage’. They “are relatively stable mental representations that represent theories about the world” and offer basic orientation in that they “guide cognitive processes like categorization and reasoning” (Evans & Green 2006: 270). This is less of a problem when we think of ‘Tuesday’ or ‘weekend’, which require an ICM of ‘week’ to make sense (cf. Lakoff 1990: 68–9), but more so with terms that are culturally loaded and are not questioned any longer, as System 1 automatically recurs to them without our conscious realisation. This is referred to as “backstage cognition” (Evans & Green 2006: 368; see also Lakoff 1990: 6) in cognitive linguistics.

In blending theory, for comparison, the term ‘mental spaces’ refers to “very partial assemblies constructed as we think and talk for purposes of local understanding and action. They contain elements and are structured by frames and cognitive models. Mental spaces are connected to long-term schematic knowledge, such as the frame for walking along a path, and to long-term specific knowledge, such as a memory of the time you climbed Mount Rainier in 2001” (Fauconnier 2007: 351). These are much more flexible and ad-hoc creations in working memory that co-exist as input spaces for on-line processes of meaning-making. When they are “entrenched” or “organized as a package we already know, we say that the mental space is framed and we call that organization a frame” ←194 | 195→(Fauconnier 2007: 352). Despite these similarities, it is significant that cognitive metaphor theory relies on frames/domains, whereas blending operates with mental spaces, which signals a more experiential and experimental approach.

Both cognitive semantics and blending theory share this focus on spontaneous, strongly contextualised meaning-making, while cognitive metaphor theory, which originally sounded as radical as enactivism today, has transformed into a much more conservative pursuit since the publication of the programmatic Metaphors We Live By. Since then it has set out to prove that conceptual mappings follow a strongly regulated pattern or grammar which has been termed the “invariance principle” (Lakoff 2007: 279; see also Evans & Green 2006: 302). It proposes that projections work in a uniform way and that metaphors form a coherent system (cf. Evans & Green 2006: 299). “Metaphorical mappings preserve the cognitive topology (that is, the image-schema structure) of the source domain, in a way consistent with the inherent structure of the target domain” (Lakoff 2007: 279). Originally, conceptual metaphor theory was very different from a schematic approach in that humans do not fill empty slots in a script or schema, but hold two or more frames in working memory, until they begin to form connections: “The essence of metaphor is understanding and experiencing one kind of thing in terms of another” (Lakoff & Johnson 2003: 5). This entails that “the only kind of similarities relevant to metaphors are experiential, not objective, similarities” (2003: 154). Since then, Lakoff’s interest has shifted from experiential contexts to a system of highly conventionalised metaphors: “Mappings should not be thought of as processes, or as algorithms that mechanically take source domain inputs and produce target domain outputs. Each mapping should be seen instead as a fixed pattern of ontological correspondences across domains” (Lakoff 2007: 275). This leads him to the claim that “metaphor resides for the most part in this huge, highly structured, fixed system. This system is anything but dead. Because it is conventional, it is used constantly and automatically with neither effort nor awareness” (2007: 293). Although conceptual metaphor theory is less helpful as a theory of reading as a process, it is indispensable to see how conceptual metaphors pervade our thinking and have become a vast network of interconnected domains that we do not even notice any longer.

A good example is the life is a journey conceptual metaphor (cf. Turner 1994: 52), which is foundational to the genre of auto/biographical writing. Lakoff explains the mappings by referencing the event structure metaphor of which life is a journey is a more specific example:

In our culture, life is assumed to be purposeful, that is, we are expected to have goals in life. In the Event Structure Metaphor, purposes are destinations and purposeful action is ←195 | 196→self-propelled motion toward a destination. A purposeful life is a long-term, purposeful activity and hence a journey. Goals in life are destinations on the journey. The actions one takes in life are self-propelled movements, and the totality of one’s actions form a path one moves along. Choosing a means to achieve a goal is choosing a path to a destination. Difficulties in life are impediments to motion. External events are large moving objects that can impede motion toward one’s life goals. One’s expected progress through life is charted in terms of a life schedule, which is conceptualized as a virtual traveler that one is expected to keep up with. (2007: 288)

Such interrelated clusters or systems are typical of conceptual metaphors. They seem natural to us due to their ubiquity and their grounding in basic human experiences or image schemas. It is hard to imagine a meaningful life that is ‘not going anywhere’, which means not pursuing a culturally acceptable goal. Failing to live up to society’s or one’s own expectations is a pervasive theme in all autobiographical writing and further complicated in illness narratives – such as cancer memoirs – where the achievement of goals is severely hampered by the impact of the disease.

A point that has not been stressed enough is the important difference between the conceptual metaphor life is a journey and the various metaphorical linguistic expressions or ‘entailments’ that are all based on this single idea: e.g. “He’s without direction in life” or “She’s gone through a lot in life” (Kövecses 2010: 3). In traditional accounts of metaphorical meaning, these surface structures would have been described as metaphorical, not the underlying concepts: “The word metaphor has come to mean a cross-domain mapping in the conceptual system. The term metaphorical expression refers to a linguistic expression (a word, phrase, or sentence) that is the surface realization of such a cross-domain mapping (this is what the word metaphor referred to in the old theory)” (Lakoff 2007: 267–8).

Conceptual metaphors, along with conceptual metonymies, play a significant role in comics narratives, as Kövecses argues, as they “are often depicted in a ‘literal’ way” (2010: 64; see also Lakoff 2007: 306), which means that abstract ideas, thoughts or emotions are externalised and literalised by visualising the image schemas on which they are based. Elisabeth El Refaie adds that in “contrast to the verbal mode, in which even the most abstract concept can, in theory, be given a verbal label, the depiction of an abstract entity in the visual mode is utterly impossible without the mediation of metaphors” (2003: 91). That explains why, for example, in an autobiographical comic the emotional toll of chronic illness has to be ‘translated’ into visual terms to make it accessible to readers (cf. El Refaie 2014: 153, 157). In the case of depression “the most common images representing the experience” are “darkness, descent, a heavy burden, or being ←196 | 197→trapped in a tight space. The process of recovery is typically framed in terms of a battle or journey” (2014: 150).

This general principle is integral to all of conceptual metaphor theory: “we typically conceptualize the nonphysical in terms of the physical – that is, we conceptualize the less clearly delineated in terms of the more clearly delineated” (Lakoff & Johnson 2003: 59; see also 105, 109, 112, 115, 248; Kövecses 2010: 7, 77). Since spatial orientation and physical interaction with objects play such an important role in embodied cognition, a lot of these basic structures become mapped through metaphors onto more complex conceptualisations: “The Invariance Principle hypothesizes that image-schema structure is always preserved by metaphor. The Invariance Principle raises the possibility 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” (Lakoff 2007: 280).

To illustrate this point, let us consider how J.R.R. Tolkien introduces Sméagol/Gollum in the second chapter of The Lord of the Rings, where a high-low, up-down orientation serves as the basis for a complex metaphorical network developed throughout the books:

There was among them a family of high repute, for it was large and wealthier than most, and it was ruled by a grandmother of the folk, stern and wise in old lore, such as they had. The most inquisitive and curious-minded of that family was called Sméagol. He was interested in roots and beginnings; he dived into deep pools; he burrowed under trees and growing plants; he tunnelled into green mounds; and he ceased to look up at the hill-tops, or the leaves on the trees, or the flowers opening in the air: his head and his eyes were downward. (Tolkien 1989: 66)

Sméagol/Gollum has a scientific mind in the tradition of Frankenstein or Strange Case of Dr. Jekyll and Mr. Hyde, two novels that heavily rely on the same metaphorical network of ‘high repute’, hubris, baseness and downfall. Here is how Victor Frankenstein describes his despicable acts: “the moon gazed on my midnight labours, while, with unrelaxed and breathless eagerness, I pursued nature to her hiding places” (Shelley 2012: 33). On top of that, we have Christianity, which is heavily invested in the same issues and adds light vs. dark and life/creation vs. death/necromancy to the network. It is no coincidence that the ‘new life’ that the villains of these novels create is based on the corruption and degradation of God’s creation: Frankenstein raises the dead, Jekyll corrupts himself, Sauron bends elves into goblins. The main point of this argument, however, is that cultures, religions and works of art employ overlapping conceptual systems based on image schemas and primary metaphors.

←197 | 198→

These primary metaphors are potentially universal, as “the major orientations up-down, in-out, central-peripheral, active-passive, etc., seem to cut across all cultures” (Lakoff & Johnson 2003: 24). However, they immediately become the basis for more complex metaphorical mappings, which means that cultures have a significant impact on metaphorical thinking: “The most fundamental values in a culture will be coherent with the metaphorical structure of the most fundamental concepts in the culture” (2003: 22; see also 14). These “partly depend on the physical environments they have developed in” (2003: 146; see also Kövecses 2010: 219), as the conceptual systems develop out of embodied cognition. Most importantly, cultures are not aware of their conceptual foundations and basic metaphors, as they are ubiquitous, deeply ingrained and seemingly natural.

That we conceive of the mind as a machine or of the visual field as a container, is something we do not realise any longer. “Ontological metaphors like these are so natural and so pervasive in our thought that they are usually taken as self-evident, direct descriptions of mental phenomena” (2003: 28). By speaking about one thing in terms of another (e.g. time is money), we begin to “understand and experience time as the kind of thing that can be spent, wasted, budgeted, invested wisely or poorly, saved, or squandered” (2003: 8). This is so ingrained in our thought processes that we do not realise that this is just one way of looking at time. System 1, of course, operates with such preconceptions all the time: “This knowledge is largely unconscious, and it is only for the purposes of analysis that we bring the mappings into awareness” (Kövecses 2010: 10; see also 72). Cultural studies, critical media literacy and several other academic/educational approaches are premised on the insight that “the way we have been brought up to perceive our world is not the only way and that it is possible to see beyond the ‘truths’ of our culture” (Lakoff & Johnson 2003: 239). This returns us to Victor Turner’s description of the rites of passage he witnessed among the Ndembu (cf. 1972: 93–111), which are an interesting example of questioning the fundamental beliefs and metaphors a culture’s thinking is based on. Literature, as Viktor Shklovsky argues (cf. 1998), attempts something similar by ‘enstranging’ the familiar.

All of this metaphorical work is ultimately aimed at making sense of the chaotic, unknowable and fluid reality we have to face: “Human purposes typically require us to impose artificial boundaries that make physical phenomena discrete just as we are: entities bounded by a surface” (Lakoff & Johnson 2003: 25). We have to reduce or expand the scale of everything to match the human sphere of experience, as we can only grasp what is accessible to us in concrete terms (cf. 2003: 34). This requires, for example, personification (cf. 2003: 33–4) to introduce active agents where there are none, e.g. in the case of cancer, to make ←198 | 199→the conceptual metaphor events are actions work (cf. Kövecses 2010: 56). Since causal relationships are fundamental to our thinking, their absence is so unbearable to us that we rather invent entities that are the source of inexplicable occurrences instead of accepting that the universe produces unforeseeable, random events. This way, the weather and its phenomena, illnesses, abstract concepts, inanimate objects etc. become personified and endowed with human qualities, so that they can perform intentional actions.

Lakoff and Johnson argue that conceptual metaphors are not tied to language alone, but to all modes of perception and expression (cf. Kövecses 2010: 12; see also 63). In this context, the role of art is said to “provide new ways of structuring our experience in terms of these natural dimensions. Works of art provide new experiential gestalts and, therefore, new coherences” (Lakoff & Johnson 2003: 235). These “emergent metaphors and emergent concepts” (2003: 58) are significant, as they allow for new experiences to be generated by bringing frames together that produce tensions and gaps. This mirrors Iser’s concept of meaning-making during reading, but extends it to multimodal texts, such as comics.

Since emotions play such a central role in aesthetic reading, it seems appropriate to finish this overview with a look at how they can be conceptualised and rationalised with the help of cognitive linguistics. The mechanism is exactly the same as with other abstract notions: they are “primarily understood by means of conceptual metaphors” (Kövecses 2010: 23; see also 380). Physical experiences (e.g. states, motions, orientations), but especially the vast field of force dynamics (cf. Kövecses 2007: 383), help us to make sense of their elusive qualities. Kövecses argues that there is a “master metaphor” (2007: 385) that provides structure and coherence to more specific conceptual metaphors: emotions are natural/physical forces. If we take anger as an example, the idea that anger is a hot fluid in a container is a more specific variant of this basic principle. It also illustrates the fact that “heat-related words account for a large portion of all the expressions that are used to talk about anger in present-day English” (2007: 394). Cognitive linguists observe that this is not a coincidence, but directly related to embodied cognition: “we subjectively experience our bodies as containers, we have the experience of a fluid inside the body, we experience heat or lack of heat in certain parts of the body, we also feel pressure when angry, and so on” (Kövecses 2010: 126). In this sense, conceptual metaphors are tied to a “folk model of emotion” that includes observations about “certain physiological effects. Thus, anger can be said to result in increased subjective body heat” (2010: 184) and “is grounded in the experience that the angry person feels ‘hot’ ” (2010: 81). Since our episodic memory stores emotions together with embodied experiences, we can easily relate postures etc. to certain emotions through metonymy, which ←199 | 200→means that “emotions can be, and are, comprehended via both their assumed typical causes and their assumed typical effects”, e.g. “shame is a decrease in size” (Kövecses 2007: 386). These are metonymical relationships in that a body posture, which is closely associated with a feeling, stands in for the feeling itself, although it is an effect of the feeling. This way, metonymies can be understood as the source of many metaphorical expressions, as the physical responses – e.g. perceived heat – then become the basis for metaphors, e.g. boiling with anger (cf. 2007: 382). These conceptual metaphors and metonymies are so ubiquitous and conventionalised in cartooning and comics that body postures as expressions of feelings are practically ‘lexicalised’ in the language of comics. Next to metaphor and metonymy, cultural models and related concepts (cf. 2007: 380), which are ideas about love, marriage, friendship etc., have an equally important influence, which, despite a shared physiological basis, makes emotion metaphors slightly different from one culture to the next.

3.5.3 Metonymies

It is now time to address the role of conceptual metonymies more systematically. Günter Radden and Zoltán Kövecses argue that “metonymy, like metaphor, is part of our everyday way of thinking, is grounded in our experience, is subject to general and systematic principles, and structures our thoughts and actions” (2007: 335). In contrast to metaphor, it is a “cognitive process in which one conceptual entity, the vehicle, provides mental access to another conceptual entity, the target, within the same idealized cognitive model” (2007: 336). Some metonymies are so entrenched that, for example, using the place instead of the institution – “The White House” for the administration of the United States, “Buckingham Palace” for the Royal Family, “Downing Street” for the British government or “Wall Street” for the financial sector in NYC (cf. Lakoff 1990: 77–8) – has become completely natural. We find entrenched metonymic relations in all conceptual categorisations and ICMs; otherwise, the instant activation of related terms would not be possible. The power of metonymy lies in activating the whole context by using a single element as a trigger.

This explains why Klaus-Uwe Panther and Linda L. Thornburg speak of a “substitution theory” (2007: 237) in the sense that metonymy represents a ‘stand for’ relation and “a predominantly referential shift phenomenon within one cognitive domain” (2007: 238). This means that metonymy can be very effective in providing an entry point to a whole domain, as it “is referential in nature: it relates to the use of expressions to ‘pinpoint’ entities in order to talk about them” (Evans & Green 2006: 311; see also 312, 315; Lakoff & Johnson 2003: 36). The meaning ←200 | 201→of the source loses most of its significance and the metonymy becomes an index that points out and foregrounds something else (cf. Panther & Thornburg 2007: 242). If, for example, a character in a film wears a white (lab) coat, the costume draws little attention to itself, but identifies the wearer as a member of the medical profession. The functional utility of the coat, its original meaning and purpose, is not relevant at all. A relationship of metonymy, contiguity and proximity turns signs into conventionalised attributes of the person they are associated with. In visual narrative media, and in comics in particular, this is frequently used to express emotions or thoughts through conceptual metonymies by having the concrete stand in for the abstract, such as the physical effects for the feelings that produced them. Nervousness can be indicated through sweat drops, for example, which does not mean that the character is literally sweating in a scene. We know that this is a metonymy because the sweat drops stand for nervousness, whereas a metaphor would mean that we understand nervousness in terms of sweat, which does not make sense in this context (cf. Evans & Green 2006: 311). From personal experience we are familiar with the biological fact that a tense situation is likely to cause physical responses, such as increased transpiration, which means that cause and effect belong to the same idealized cognitive model (ICM) (cf. Evans & Green 2006: 312). Comics use metonymy as a convenient shorthand that allows for the representation of complex subject matters via radically reduced visual signs.

To illustrate the ubiquity of conceptual metonymies in everyday conversations, I would like to discuss the example ‘I speak English’ (cf. Radden & Kövecses 2007: 342), which seems so familiar that we do not question it any longer. That we accept this sentence as meaningful is rather astounding, as every one of its constituent elements raises questions. All three words, ‘I’, ‘speak’ and ‘English’, have a metonymic relation to the heterogeneous experiences that they ‘stand for’. For a non-native speaker of English who lives in a German-speaking country the ‘I’ did not speak English (well enough) for a very long time to warrant that claim and then only under very specific circumstances, e.g. as a teacher, as a tourist etc. The ‘English’ I spoke as a ten-year old was very different from the English I used in grammar school, during my year abroad twenty years ago, and the English I speak in class today etc. So both ‘I’ and ‘English’ stand in as wholes for the more specific parts they refer to. This is the generic for specific metonymy like “Boys don’t cry” (Radden & Kövecses 2007: 343), which we usually call a ‘generalisation’.

There are further metonymies present: ‘speaking’ may be the signature skill of language learning, but it has to stand in for a substantial range of knowledge structures, skills and competences that are necessary, but only implied ←201 | 202→here, at best. It also expresses an ability or potential, something I could do, as a fact by foregrounding the active skill that best proves that the sentence is true, which is being able to speak rather than being able to read. The truth value of this statement cannot be expressed through modal verbs, which would immediately raise doubts: I could speak English. This is based on the actual for potential or immediate over non-immediate metonymies that also apply to constructions like “He is very helpful” (cf. Radden & Kövecses 2007: 343). In all these examples, what is expressed as a claim about the present – without providing any direct proof – relies on past events which become blended into a generic, highly underspecified statement that provides clues hinting at richer contexts of meaning, but do not mean a lot in and of themselves.

A similar thing happens when we identify our cars in a car park or ourselves in a group photo by stating: “That’s me!” We are not aware that this is a conceptual metonymy and that it requires a somewhat broad understanding of our personal identities that include tools and visual representations. In the second case, it is easier to find a rational explanation why we are willing to identify people via their faces: “We derive the basic information about a person from the person’s face. The conceptual metonymy the face for the person is part of our everyday way of thinking about people” (Radden & Kövecses 2007: 335). This is not only relevant in terms of personal identification and daily conversations, but also to the ways in which visual narrative media grant access to characters’ inner states through facial expressions. “From a semiotic perspective, metonymy is related to indexicality” (Panther & Thornburg 2007: 242) in that it functions as a pointer to otherwise inaccessible contexts.

Apart from such conceptual metonymies, Radden and Kövecses also include basic semiotic relationships, such as conventional signs and symbols, under the umbrella term of metonymy, although these links are arbitrary and culturally determined (cf. 2007: 338). Here it is hard to draw a line between elements that naturally or culturally co-occur, provided that the distinction is possible at all. Is it more natural to associate sails with boats than to link a company’s logo to its name and products? In both cases the two elements belong to the same ICM, which means that, cognitively, there is no difference. In the case of ‘[[insert_content]]#x2019; or ‘dollar’ for ‘money’, “the form metonymically stands for the concept it denotes” (Radden & Kövecses 2007: 338). These signs and concepts, then, have a metonymic relationship to reality in the form of ‘reference metonymies’. This approach would make all of language metonymic, as it claims that both thought and language have a shared metonymic basis.

An interesting claim that Kövecses puts forward is that “many conceptual metaphors have a metonymic basis or motivation” (Kövecses 2010: 185). This ←202 | 203→is another grey area in terms of how far one is willing to stretch the concept. If all of our knowledge is stored in frames/domains/ICMs and all the links within them are metonymic, then many conceptual metaphors – in their entrenched forms – could also be understood as metonymies, as they are not mapping structure between domains but are conceptually bound to the experience of a single domain. Considering that the feeling of anger often co-occurs with a red face, sweat, heat, pressure, a need to control oneself to avoid any regrettable actions etc., it is feasible to argue that they form a coherent experience and thus a single domain. The metaphor would cease to exist, as soon as the mappings become fully integrated into a single domain. That is why blending theory distinguishes between cognitive processes and their mental or material products.

3.5.4 Blending

The most comprehensive introduction to conceptual blending theory can be found in Gilles Fauconnier and Mark Turner’s The Way We Think (2003). It adheres to the same tenets as all strands of cognitive linguistics, especially in its reorientation away from surface linguistic expressions to concepts, and here again from conceptual spaces to their networked interrelations. As in cognitive semantics, the importance of language is reduced to that of prompts (cf. 2003: 142–3, 146–7, 183, 277, 360). A further sign of continuity is Fauconnier and Turner’s integration of conceptual metaphors into blending theory as one of its four basic types (cf. 2003: 127). Evans and Green try to alleviate potential tensions between the proponents of both theories by claiming that they are, in fact, complementary (cf. Evans & Green 2006: 435) and may fulfil different functions. While in conceptual metaphor theory the focus has shifted to a diachronic exploration of how the vast system of metaphors has come into existence, relying on systematic rules and conventional domains, conceptual blending theory is more concerned with meaning-making as a creative act. This is not to imply that blending theory has nothing to say about fossilised integration networks or complex cultural blends, but that it lends itself more easily to a conceptualisation of reading as an ongoing cognitive process. This becomes apparent in blending theory’s reliance on mental spaces. Fauconnier and Turner state that these “are built up dynamically in working memory, but they can also become entrenched in long-term memory. For example, frames are entrenched mental spaces that we can activate all at once” (2003: 103; see also Fauconnier 2007: 365). It is important to understand that most mental spaces are “the products of blending” (Fauconnier & Turner 2003: 104) and may, in turn, be reactivated and reblended (cf. 2003: 24, 279), much in the same way ←203 | 204→as Iser describes perpetual gestalt-forming. While there is a clear purpose and teleological drive in Iser’s theory of reading, Fauconnier and Turner understand blending as a basic cognitive mechanism: “Conceptual work is never-ending” (2008: 61). They discuss four types, two of which we have already encountered.

Filling the slots in mental models is the first and most basic blending phenomenon, which Fauconnier and Turner call a ‘simplex network’ (cf. Fauconnier & Turner 2003: 120–2). Here we have two spaces, a frame with roles, such as a form, and elements/values that fit these roles, such as personal data. By filling in the form we create a blend between the generic and the specific. The structural frame of the first input space is inflexible and determines the selection from the second input space.

The second type of blending is a ‘mirror network’ (cf. 2003: 122–6), which operates with at least four mental spaces. The two new ones are the generic space and the blended space. The generic space serves as a schema or “template” (Evans & Green 2006: 406) that indicates what we are looking for across the input spaces. In this way, its organising frame “provides a topology” that “specifies the nature of the relevant activity, events, and participants” (Fauconnier & Turner 2003: 123) across all mental spaces involved in the network (cf. 2003: 122). Such a blend is easy to establish as the things we are looking for mirror each other: “Elements in the generic space are mapped onto counterparts in each of the input spaces, which motivates the identification of cross-space counterparts in the input spaces” (Evans & Green 2006: 404). A simple example is the blending of repeated experiences with slight variations into a single frame – such as a morning routine. The generic frame provides some structure when we look for common patterns across several days. We ignore all the information that is irrelevant for this search (e.g. outdoor temperature, type of breakfast) and only focus on things in common, such as the typical location, average times and the usual sequence of activities. This explains how prototypes, templates and generic models are created in the first place, which in turn serve as structures for generic spaces in future blends (cf. 2003: 116).

The third type is called a ‘single-scope network’ (cf. 2003: 126–31) and corresponds to the mappings of conceptual metaphors (cf. 2003: 127). Both input spaces have separate and potentially contradictory organising frames, but only one of them is used to organise information in the other. Therefore, projection is unidirectional (cf. 2003: 126) and circumvents potential contradictions by highlighting similar structures and ignoring elements that do not fit. Still, one cannot fail to notice “a highly visible type of conceptual clash” (2003: 129), as the organising frame of the target space is strategically ignored. With the time is money metaphor, for example, only a select few structures are mapped from ←204 | 205→‘money’ onto ‘time’, for which we have to ignore a lot of the specific characteristics of the latter to make the metaphor work. The generic space emerges through cross-space mappings between the inputs and guides this process in turn. The blend represents the metaphor as an active network between the input spaces.

Fauconnier and Turner refer to the fourth type as a ‘double-scope network’ (cf. 2003: 131–5), which can be considered as the most complex type of blending. In this configuration all input spaces contribute elements to the blend and project their organising frames: “A double-scope network has inputs with different (and often clashing) organizing frames as well as an organizing frame for the blend that includes parts of each of those frames and has emergent structure of its own. In such networks, both organizing frames make central contributions to the blend, and their sharp differences offer the possibility of rich clashes” (2003: 131). Double-scope blending is the unique human ability to overcome differences with a shared vision: “conceptual integration networks offer a way to see unity behind the diversity of particular manifestations of meaning constructions” (2003: 137). Like Iser’s model, it involves a ‘seeing together’ that favours the correspondences over the divisive matter. The ‘emergent structure’ or meaning is unique to the blend and a result of integrating several mental spaces at once (cf. 2003: 42). Complex blends are preliminary and tentative at first and may require further input to work.

There are three processes that facilitate “the creation of new meaning in the blend” (2003: 20): (1) composition refers to the establishment of new relations, combinations and the integration of different aspects; (2) completion involves the activation of additional cognitive frames and input spaces to add further structure and background knowledge, mainly to stabilise the blend; (3) elaboration or dynamically ‘running the blend’ allows for the emergence of new meanings by thinking further along the lines that have been established (cf. Evans & Green 2006: 409–10). When the network ‘gels’, it leads to what Fauconnier and Turner call ‘global insight’ or – to be more precise – “the impression of global insight” (2003: 323). This is a personal eureka moment that may be based on a System 1 blend, a sudden realisation out of the blue, or the result of running the blend consciously for some time (System 2).

What we call inspiration, creativity or epiphany are usually emergent meanings in a double-scope blend: “The products of conceptual blending are always imaginative and creative” (2003: 6). According to conceptual blending theory, creativity is not the invention of completely new things, but the imaginative recombination or modification of existing material (cf. 2003: 146–7; Evans & Green 2006: 400–1). Global insight energises and illuminates the entire network. The input spaces are not replaced by the blend, but they may be transformed ←205 | 206→in light of the new insights: “The integration of events in the blend is indexed to events in both of the input spaces. We know how to translate structure in the blend back to structure in the inputs. The blend is an integrated platform for organizing and developing those other spaces” (Fauconnier & Turner 2007: 377). This process is known as “backward projection” (Evans & Green 2006: 410; see also Fauconnier & Turner 2007: 366) and is conceptualised in the following way: “As we project to a blend, we are also working on the entire network, and we may, for instance, recruit new structure to the inputs precisely to make it available for possible projection to the blend. […] Input formation, projection, completion, and elaboration all go on at the same time, and a lot of conceptual scaffolding goes up that we never see in the final result” (Fauconnier & Turner 2003: 72; see also 94, 129). Since “the connections between the blend and the inputs remain active, applying imagination to the blend has consequences for the inputs” (2003: 60–1; see also 44), which means that spaces “can be modified at any moment in the construction of the integration network. For example, the inputs can be modified by reverse mapping from the blend” (2003: 49).

Fauconnier and Turner claim that blending does not represent an unusual phenomenon or that it only occurs in artistic production, but represents a standard, largely automated cognitive process: “human beings are exceptionally adept at integrating two extraordinarily different inputs to create new emergent structures, which result in new tools, new technologies, and new ways of thinking” (2003: 27; see also 317). To facilitate this ease of conceptual integration, “blending operates largely behind the scenes. We are not consciously aware of its hidden complexities, any more than we are consciously aware of the complexities of perception” (2003: v; see also 5, 12, 14, 18, 33–4, 71). Like Iser (cf. 1980) or Kahneman (cf. 2012), they argue that consistency-building is largely automatic, but that we are, of course, aware of “the products of blending” (2003: 391; see also 57, 78) and can ‘run the blend’ actively: “Nearly all important thinking takes place outside of consciousness and is not available on introspection; the mental feats we think of as the most impressive are trivial compared to everyday capacities; the imagination is always at work in ways that consciousness does not apprehend; consciousness can glimpse only a few vestiges of what the mind is doing” (2003: 33–4). However, relying on System 1 comes with its own problems, as we have already seen with Kahneman:

Composition, completion, and elaboration all recruit selectively from our most favored patterns of knowing and thinking. This makes blending a powerful cognitive instrument, but it also makes it highly subject to bias. Composition, completion, and elaboration operate for the most part automatically and below the horizon of conscious observation. This makes the detection of biases difficult. Seepage into the blend can ←206 | 207→come from defaults, prototypes, category information, conventional scenarios, and any other routine knowledge. (Fauconnier & Turner 2007: 392)

Since social media and our own habits only provide us with inputs that do not require more than a mirror network, using our own prejudices for guidance in the generic space, so that we can easily identify and blend the patterns we already know and prefer, we desperately need the double-scope blending of a democratic debate in which each person gets to project his or her own organising frame to the blended space. This is why Werner Delanoy speaks of dialogue as the essence of teaching (cf. Delanoy 2002, 2008). Of course, this may lead to a “highly visible type of conceptual clash, since the inputs have different frames” (Fauconnier & Turner 2003: 129), but classroom discussion, as I have argued before, is essential in our society to learn how to co-create meaning through interaction. Teachers have to understand that their attempts at single-scope blending, for which one input space provides the organising frame for all the other input spaces (students), is not conducive to such an endeavour.

This has been an example of running an ad-hoc blend by bringing two distinct frames, conceptual blending theory and classroom discussions, together. The success or failure of such an attempt is determined by how appealing the emergent meaning of the blend is in contrast to the clash of structures, such as comparing mental processes to human interaction. Blends are never perfect and never finished. If the current state of this network is not appealing enough, we could run the blend a few more times and see if we can recruit structures from classroom discourse to illuminate how blending works. The important thing to realise here is that both Iser and Fauconnier/Turner claim that meaning-making works exactly like this: we keep at least two distinct input spaces in mind for mutual illumination and begin to see both in terms of each other.

We have already encountered double-scope networks, such as in Miall and Kuiken’s theory of border-crossing and self-modifying feelings. Here the input spaces are, on the one hand, readers’ memories, thoughts and feelings and, on the other, a character’s unique circumstances (cf. Schneider 2012: 17). This leads to what Ted Cohen calls “imagining some third person, some new person, some blend of what I know of you and what I know of me” (1999: 402). Provided that new meaning emerges in the blend, self-modifying feelings have repercussions on the input spaces through “backward projection” (Evans & Green 2006: 410). Fauconnier and Turner speak of “the impression of global insight” (2003: 323), as the blend may also lead to wrong conclusions and the appropriation of characters. It is now easier to understand why some critics find the ‘metaphor of personal identification’ potentially problematic in the context of empathy. If ←207 | 208→we understand the concept as a single-scope blend, it would mean that we project the organising frame of our own experiences onto the perceived life of a character, which would negate incongruous elements in favour of a harmonious blend, as we disregard specific circumstances that are incompatible with our own experiences. Since blending is usually a System 1 operation, the flow of reading can facilitate a superficial identification that makes us believe to be very close to a character and understand exactly how he or she feels. Or we conceptualise empathy as a double-scope blend, which I strongly suggest, but then there is no merging and complete identification. We still get the vital relations and emergent meanings – or ‘fresh feelings’, as Miall and Kuiken call the result – which affect both our understanding of the character and ourselves. At the same time, we are aware of the differences and see the undeniable importance of the text – the special insight – as resulting from a partial overlap or shared humanity rather than a perfect match of personalities. Since this is a very unusual context, I continue with a more relatable example.

Most personal computers still have a ‘desktop’, which is a computer interface that simulates an arrangement of typical office equipment, such as a calendar, a notepad, a wastepaper basket, folders, a calculator etc. as if they were spatially arranged on top of a desk (cf. 2003: 22–24, 131). Thus, a so-called desktop computer runs a simulation of the actual desktop on which it is placed, mainly to camouflage its alien presence and to make it more user-friendly. We have two input spaces with their own internal logic and organising frames: the domain of traditional office work and computers with their unique input devices, such as a keyboard and a mouse. The first frame provides familiar objects, but also typical activities: adding dates to calendars, taking notes, opening folders, looking for documents, throwing away stuff into the bin etc. In contrast, the computer requires the user to handle a mouse successfully (navigating, pointing, scrolling, clicking), to understand its organisational logic of ‘windows’, to navigate pop-up and pull-down menus, but also to click icons/buttons to get things done. When computers entered the workspace, all of this had nothing to do with the objects and activities in the other input space. The conceptual integration of these two frames required a lot of imagination. Importantly, the blend “is not the screen: The blend is an imaginative mental creation that lets us use the computer hardware and software effectively” (2003: 23). We still do office work, but in a novel way. We open folders by double-clicking them, we add files to folders by dragging one icon onto another instead of physically adding sheets of paper to a folder, we add notes or additional information to a document via a keyboard and save this new version over the previous one etc. There are, of course, also discordant elements – as with every double-scope blend – that we do not notice any ←208 | 209→longer: the wastepaper basket sits on the desktop and has exactly the same size as all the other objects; all elements are aligned in horizontal and vertical lines; the content of folders is displayed as pop-up windows; and, most importantly of all, all these operations do not actually take place and none of the objects really exist. There are no calendars, photos, clocks, calculators, folders, desktops, notebooks, telephones etc. anywhere in the system. Since there is no physical desktop in the first place, nothing can be ‘on’ it. That all of this seems natural to us has a lot to do with the successful blends that are involved: as long as they work for us on a human scale, we are more than willing to live inside the blend. The mouse simulates a reaching of the hand, the double click opens the folder, the pop-up window equals the movement of bringing a sheet of paper closer to the eyes, the layout of text on the simulated page is made to look exactly like the real thing etc. In a line from hand writing to typewriters to desktop computers to tablets with styluses we find an ongoing attempt to create a perfect blend between two otherwise very different frames.

If blending is closely associated with the imagination and requires the active engagement of creative minds, the question arises whether blending could be too subjective, unpredictable and ephemeral. However, blends have already proven their worth in the form of scientific theories, religions, works of art, cultural artifacts, rituals and physical objects: “Cultures work hard to develop integration resources that can then be handed on with relative ease” (2003: 72; see also 321), which means that “cultural concepts are the products of successive blending over generations” (2003: 295). A second answer has to do with the cognitive operation of blending itself, which is far less obscure than Iser’s gestalt-forming: “conceptual blending is a general, basic mental operation with highly elaborate dynamic principles and governing constraints” (2003: 37; see also 17, 29, 168). Any type of map or timeline, for example, from (human) evolution to world history, is a blend that compresses vast amounts of data into a simple list or drawing that fits onto a book page and can be comprehended in a fairly short amount of time.

There are a number of elements – or vital relations – that regularly become compressed in blends: time, space, identity, cause-effect, part-whole (metonymy), change, representation, role-value, analogy, property, (dis)analogy, similarity, category and intentionality (events are actions) (cf. 2003: 93–102). This sounds more complicated than it is, which can be illustrated with a simple example: the evolution of human beings has involved millions of individuals over millions of years, but when we watch a documentary on TV, we see one individual transform from what looks like an ape into a modern human individual in about a minute through computer animation. This is what Fauconnier and Turner call compression into uniqueness: millions of individuals become one ←209 | 210→being, millions of years one minute, and a variety of locations all across the globe a single generic setting. In conceptual integration theory this is a “compression of ‘outer-space’ links into ‘inner-space’ relations under blending” (2003: 93). In other words: the numerous cross-space mappings, which are ‘outer-space’ from the point of view of the input spaces, become condensed into a single relation within the blended space itself. This is important, as we can store and retrieve coherent frames as single units – especially when they follow the subgoals of compression that Fauconnier and Turner have defined: “Compress what is diffuse. Obtain global insight. Strengthen vital relations. Come up with a story. Go from Many to One” (2003: 312).

This simplification of and disregard for elements that do not fit come with their own problems, but unfortunately we live by these metaphors and blends, no matter if we want to or not: “Human beings are evolved and culturally supported to deal with reality at human scale – that is, through direct action and perception inside familiar frames, typically involving few participants and direct intentionality” (2003: 322; see also 324). Through scaling, the shortening of time (syncopation), monocausal explanations, the events are actions metaphor and quick, dramatic changes, a real human life can become a hagiographic text, for example: “We can compress a lifetime not only by scaling it to run very fast but also by dropping out all but a few key moments (being born, meeting Christ, being shot through with arrows, going to heaven). Scaling and syncopation often work together” (2003: 314).

The Gothic cathedral is a great example of how a complex doctrine was literalised and materialised in the form of a building and turned into an experience for a largely illiterate group of church-goers (cf. 2003: 207–10). Based on perpetual blending, the Church was able to compress, for example, narratives from the Bible into single paintings, hagiographies into statues, the relationship of the congregation with God into architecture etc. These serve as material anchors (cf. Keen 2010: 67; Kövecses 2010: 279–81; Oatley 2013: 452) or encapsulated blends, as it were, that help to decompress the elaborate narratives on which they are based, activating the corresponding frames through metonymy. The rosary is another good example, as it compresses a very long chain of interconnected prayers into a physical chain with wooden beads that serves as a mnemonic device. However, there are also countless material anchors without a religious context, such as maps, rings or watches (cf. Fauconnier & Turner 2003: 195–8, 332) that are equally successful. Edwin Hutchins argues that the “ability to combine conceptual structure with material structure is a key cognitive strategy” (2005: 1556), which also extends to social practices, such as queuing:

←210 | 211→

Consider a line of people queuing for theatre tickets. This cultural practice creates a spatial memory for the order of arrival of clients. The participants use their own bodies and the locations of their bodies in space to encode order relations. The gestalt principle of linearity makes the line configuration perceptually salient. Our perceptual systems have a natural bias to find line-like structure. But seeing a line is not sufficient to make a queue. Not all lines are queues. Soldiers standing at attention in formation form a line, but not a queue. In order to see a line as a queue, one must project conceptual structure onto the line. (2005: 1559)

In this sense, a “physical structure is not a material anchor because of some intrinsic quality, but because of the way it is used” (2005: 1562). Following this logic, it becomes obvious how the human mind seeks material crutches in the environment that constitute a far-reaching support network of which we are no longer aware. Some also believe that they can gain access to other people – dead or alive – by visiting places or handling objects that are associated with them, which explains all cultures of remembrance:

This use of space as a prompt to blend events, intentionality, and times is a basic cultural instrument: We visit the graves of dead relatives, heroes, and martyrs; we visit the towns where Vermeer and Shakespeare were born; we return to our alma mater; we go to chapels or churches to pray even when there is no service, and of course the graves are either in the floor of the church or in the graveyard next to the church. Part of the motivation for these visits is the sense that, if we actually inhabit them, we can more easily integrate our thinking and emotions with the people, cultures, and events associated with them, no matter how ancient. (Fauconnier & Turner 2003: 316)

Fauconnier and Turner stress the fact that “we need always to keep in mind the distinction between the operation of conceptual blending and the cultural products of conceptual blending” (2003: 215). Both are based on the same principles, but cultural artefacts are successful blends that have become entrenched and materialised, whereas blending as a process is a standard cognitive operation. Androids as blends of humans and machines are interesting and highly successful material anchors in science fiction literature, as they allow us to think through our own humanity and identity. The function of narrative, in this case, is to decompress the blend and explore the structures that have not been projected from the inputs due to their incompatibility.

The same can be said about life writing. We tend to forget that our own identities are blends (cf. 2003: 118), which Fauconnier and Turner explain in the following manner:

Personal identity itself involves a diffuse network of mental spaces whose compression in the blend creates the unique person. Conceptually, a person is involved in mental spaces over many times and places, through many changes. All those spaces contribute ←211 | 212→to a blend that has the single unique person. There is a physical material anchor for this conceptual blend – the active living biological body that we can see and with which we can interact. We can hear its voice, and it can hear ours. When the person dies, the conceptual network with the unique person persists for us, if not for the person. (2003: 205; see also 95)

This, of course, is going to be a major concern of part 5, together with material anchors. By blending our experiences with other people across different encounters, “we are able to extract regularities over different behaviors by the same person to build up a generic space for that person – a personal character” (2003: 251–2). This observation illustrates the difference between a narratological and an aesthetic approach to literature: narratologists are interested in the generic spaces of the framework and their mappings, whereas aesthetic reading looks at the input spaces and blends. This is not to deny that both exist within the same framework and contribute to reading in a meaningful way: generic spaces provide global structure and orientation, but the imaginative work of reading takes place somewhere else.

Despite the fact that conceptual integration theory (blending) was never intended to be understood as a theory of reading, its general principles and cognitive operations share a number of characteristics with reader-response criticism. Louise M. Rosenblatt’s book, for example, carries two ‘input spaces’ and the blend in the title: The Reader, the Text, the Poem. In Teaching Literature: Nine to Fourteen Benton and Fox define four input spaces – the new reading text, previous literary experiences, personal experiences and world knowledge, which can be blended into a reader’s unique understanding of a text. Significantly, they describe conceptual integration as a form of creative work:

In the act of creating, what the reader brings to a story is as important as what the text offers in the sense that we fit the reading of a new story into the blend of our literary and life experiences to date, drawing upon our knowledge of other fictions as well as upon analogies in the primary world, in order to make our own, unique meaning. (1985: 5)

The idea of the reading text as an ‘input space’, however, has to be rejected for a number of reasons: it is too complex for a single frame, it exists outside of human cognition as a material object in the world and, following the logic of reader-response criticism, it is a mere ‘blueprint’ that requires a human mind’s active engagement with it. However, this third point also offers a solution as to how blending theory and reader-response criticism can be brought in line. Literary texts prompt the construction of tentative mental spaces – gestalten – that are already blends, as our first impressions of a character, for example, heavily depend on a whole range of previous experiences with literary texts, ←212 | 213→our personal memories and knowledge about the world etc. Much more importantly, the narrative immediately starts to invite cross-space mappings and further blends. Iser calls these prompts “strategies” (1980: 96), which encourage blending by foregrounding structures that are relevant to the narrative’s stage of development and overarching themes. These correspondences between conceptual integration theory and reader-response criticism are addressed in the next chapter, especially because they have not been sufficiently acknowledged yet.

3.6 Blending & Literary Studies

There are at least two major publications that propose an application of blending theory to reading literature: Barbara Dancygier’s monograph The Language of Stories and Ralf Schneider and Marcus Hartner’s edited volume Blending and the Study of Narrative. Since the collection of essays covers a variety of media, narratological categories and specific case studies, I restrict my discussion to The Language of Stories. Looking at Dancygier’s general conceptualisation of reading, one is struck by literally dozens of passages that seem to be directly taken from either Wolfgang Iser or Louise M. Rosenblatt’s books. Here is Dancygier’s argument against the idea that the public meaning of a text can be arrived at through one specific approach: “while meaning is not entirely indeterminate, it is also not determinate. It is a perfectly natural reaction on the part of those engaged with various sources of interpretation to feel that someone arguing for a single interpretation is missing the point” (2012: 9). She goes on to argue that “there is an impressive number of possible readings of the story all dependent on how the reader (not the text) construes the spaces set up” (2012: 38), so that “whatever understanding a reader might acquire, it is not contained ‘in’ the story, but can only be arrived at through the interaction with it” (2012: 203). This is Rosenblatt’s transactional theory in a nutshell. Astonishingly, Dancygier neither read Iser’s The Act of Reading nor Rosenblatt’s The Reader, the Text, the Poem, judging from the absence of both titles from the bibliography (cf. 2012: 216, 219). She does, however, list Iser’s The Implied Reader for one brief reference only (cf. 2012: 58, 216).

Such a noticeable disregard for reader-response criticism can be found across all recent cognitive approaches to literature. A laudable exception is Marco Caracciolo, who extensively credits Iser (cf. e.g. 2014: 12–13, 36, 64, 68, 77) and even traces his own approach back to John Dewey (cf. 2014: 22–4, 51, 72–90). One potential reason for this discontinuity could be the substantial backlash against Iser’s idiosyncratic conceptualisation of the reading process since the heyday of reader-response criticism in the late 1970s and early 1980s. Maria Bortolussi ←213 | 214→and Peter Dixon (cf. 2003: 5–7), for example, condemn what they understand as an “indulgence in circular logic, speculative hypothesis, capricious use of terminology, and monolithic views of reading experience” (Bortolussi & Dixon 2003: 6). Iser’s work is ultimately worthless, they imply, because his arguments are not based on empirical research:

Although Iser’s intuitive descriptions of the reading process provide some interesting insights, they remain purely speculative because his text-based approach offers no method of validating the hypotheses. Consequently, his theory sheds little light on what actually transpires in the mind of readers during the reading process. (Bortolussi & Dixon 2003: 7)

Whether narrative psychologists should have the last word on ‘what actually transpires in the mind of readers’ has already been answered. Instead of discrediting Iser, it may be worth acknowledging the ‘interesting insights’ that anticipated some of the central concepts of blending theory. What he referred to as an image or a gestalt, is now called a blend. His ‘retroactive effect’ corresponds to ‘backward projection’, which Dancygier explains by stating that “subsequent blends throughout the process affect the understanding of the inputs” (2012: 56). Here is Iser’s explanation of this phenomenon:

In most literary texts, however, the sequence of sentences is so structured that the correlates serve to modify and even frustrate the expectations they have aroused. In so doing, they automatically have a retroactive effect on what has already been read, which now appears quite different. Furthermore, what has been read shrinks in the memory to a foreshortened background, but it is being constantly evoked in a new context and so modified by new correlates that instigate a restructuring of past syntheses. (1980: 111)

Iser’s terminology (‘foreshortened background’, ‘correlates’ or ‘syntheses’) clearly describes blending phenomena, even if he did not have the means to explain them in such a systematic fashion as Fauconnier and Turner managed to do. His idea of gradual gestalt-forming is directly mirrored in Dancygier’s following argument: “Similar meaning construction processes occur in longer narratives. Frames and mental spaces structure inputs, which then become integrated, possibly in ways specific to a reader, into the emergent blend. The process continues throughout reading, until the complete blend of the story emerges” (2012: 35). Here is a similar description relating to the final gestalt, which provides superior closure:

A ‘story’ can thus be discussed as a cognitive construct, a blend, emerging through the process of meaning construction triggered by reading. The role of the text is crucial in providing such prompts, but the emergence of the story relies to a comparable degree on the frames evoked in the reader’s mind, and on the construction of double-scope blends, ←214 | 215→integrated into the mega-blend. The story is the mega-blend arrived at in the interaction with the text. (2012: 56)

The following quotation explains Iser’s theme-horizon structure: “All these are mental spaces, activated for the duration of this part of the conversation, and then becoming latent until evoked again” (2012: 35). Both theories rely on the cognitive management of several mental spaces at once that may become reactivated much later in the narrative. In part 1 I argued that Iser’s theory of gestalt-forming transcends the linearity of the narrative and dismisses the idea that, at the end of the narrative, a mega-blend provides global insight. Dancygier comes to the same conclusion: “The process of reading is thus not simply linear and does not rely primarily on the accumulation of information. It is a multidimensional process, reaching across narrative spaces in different directions” (2012: 197). These cross-space mappings are encouraged by “referential links [that] may be established not only backwards and forwards in the flow of discourse, but also across spaces” (2012: 118). When Dancygier introduces her concept of ‘narrative spaces’, which are the equivalents of mental spaces in the context of reading, she highlights the fact that they are developed throughout the reading process and are potentially relevant to all scenes:

A narrative space is thus a construct which is set up through linguistic means and continues being elaborated through some parts of the text (possibly all). It is also subsequently enriched through blending and gradually starts functioning in the network leading to an emergent story. In these respects it is thus similar to a mental space, which participates in extended discourse. What constitutes a crucial difference, however, is the nature of the discourse, since a narrative requires that its primary spaces be maintained and elaborated until the completion of the reading process, until their role in the text is fulfilled. (2012: 37)

In other words: the retrieval of episodic memories (previous scenes) from long-term memory, but also their re-evaluation and reblending play a much larger role. Literary texts use foregrounding very effectively to invite cross-space mappings between current blends and previously formed gestalten. Iser’s “response-inviting structures, which impel the reader to grasp the text” (1980: 34) become prompts in blending theory (cf. Dancygier 2012: 54) or ‘narrative anchors’ that provide metonymic ties to narrative spaces (cf. 2012: 49):

I argued that the fragmented nature of most narratives requires specific narrative mechanisms which provide coherence links across different narrative spaces. I defined narrative anchors as expressions which set up or suggest the availability of narrative spaces, but do not elaborate them right away. Such ‘place-holders’ may activate new narrative spaces and allow them to remain active, but the spaces are elaborated gradually as the text unfolds, and often contribute to the topology of other space constituting ←215 | 216→the story. The second function of anchors is ‘link-building’. The links they set up may prompt what I have called cross-input projection – spreading of topology from one input to another and building the coherence and completeness of the emergent story. (2012: 42)

This means that the “emerging complex referential networks” (2012: 117) offer many different ways of connecting the dots. These ‘dots’ are “narrative anchors” or “narratively salient expressions which rely metonymically on frames and exercise our representational abilities” (2012: 50). This may happen, for example, through “repetition of some of that information” (2012: 43). Dancygier provides an example from Margaret Atwood’s The Blind Assassin that foregrounds how narrative anchoring can be strategically used by authors to invite specific types of cross-space mapping:

Another type of anchor consists in evocation and repeated re-activation. A mention of Alex as wearing a blue, worker’s shirt but smoking ready-made cigarettes is a salient descriptive detail, and its reactivation in a different narrative space prompts the cross-input projection linking Alex to ‘him’ and making this cross-space identity available in both spaces. This type of anchoring relies as much on the reader’s attentiveness and memory as on the salience of the frames thus constructed and on the sheer number of anchors establishing and reestablishing the same cross-input links. Crucially, such anchors not only link the spaces, but also prompt projections of topology from input to input (narrative space to narrative space). (2012: 44)

This is why Michael Benton promotes the conceptual metaphor readers are detectives (cf. 1992: 44). If meaning-making in reader-response criticism and blending are essentially the same mental operations, it should not be surprising that the role of readers is conceptualised in similar terms. For comics as a narrative medium it is crucial that “anchors may also be images which form an entire network of concepts and jointly give meaning to an abstract and difficult text” (Dancygier 2012: 43), which is going to be a major point in the next part, where Thierry Groensteen’s ‘braiding’ or ‘tressage’ (cf. Groensteen 2007: 145–9) – the comics term for this phenomenon – is still underestimated in favour of Scott McCloud’s strict linearity.

This brief comparison should serve as further proof that there are too many similarities between reader-response criticism and conceptual integration theory to speak of a coincidence. I would even claim that Dancygier’s approach is an updated version of reader-response criticism with a better theoretical basis and a clearer concept of how to conceptualise reading in progressive stages. Her tentative, exploratory approach is compatible with aesthetic reading, but not with data mining and narratological analysis. Blending theory is interested in meaningful links between spaces rather than in detailed lists of what they contain. ←216 | 217→While the associations and potential links often appear out of nowhere – they are generated by System 1 – we can ‘run the blend’ consciously and see how far it takes us. Contrary to the right/wrong and relevant/irrelevant dichotomies of efferent reading, blends are never perfect. They are tentative projections of what is happening in a narrative, which always means sacrificing details that do not fit into one particular reading of a text. Yet, the emergent meaning in the blended space may be intriguing enough and, as long as readers can run with it, they are bound to do exactly that. A blend/gestalt is something on which readers actively work, but it can never be the same for everyone, perfect or complete. For the remainder of this chapter the focus shifts to key ideas in Dancygier’s approach that are either only partly addressed in reader-response criticism or not at all. These include ‘compression’ and ‘decompression’ as artistic choices and their interplay with the same processes performed by readers, her theory of focalisation and viewpoint compression, her interest in enactivism and the accompanying criticism of Theory of Mind, the impact of multimodality on blending and, finally, ‘fictive vision’.

Up to this point, we have mainly looked at blending as a meaning-making process that compresses several input spaces into a single blended space. However, most of the mental spaces we operate with on a daily basis are already frames: complex networks that can be retrieved all at once. This is especially true of literature, where writers/creators may rely on their readers’ ability to handle such blends with ease. At the same time, they may choose to decompress complex structures to allow readers to discover the connections for themselves. Accordingly, Dancygier speaks of these two processes as artistic choices and narrative strategies. The first is familiar from reader-response criticism. Here is John Dewey’s explanation again: “For art is a selection of what is significant, with rejection by the very same impulse of what is irrelevant, and thereby the significant is compressed and intensified” (2005: 217; see also 305). Iser equally talks about selection in the context of the repertoire (cf. 1980: 82; see also 109), which makes the selected elements to stand out and become “highly determinate” (1980: 85). When Dancygier describes the “crucial effect emerging out of the compression”, she stresses the artistic control over the material that leads to “an enhancement of the central themes and increased salience” (2012: 59). Since compression is crucial to blending, Dancygier argues that it “will affect all aspects of the narrative – time, viewpoint, characterization” (2012: 59). However, intense compression runs counter to the idea that readers can vicariously experience a character’s life as it unfolds in meaningful scenes and interactions with other characters. This is especially relevant in the context of autobiographical writing, where the authors begin their journey of self-exploration with mega-blends on which they ←217 | 218→have been working for many years. These have to be decompressed first to allow for more accessibility and subtlety:

One’s sense of uniqueness is a result of a highly compressed blend, but it is natural to decompress that whole when need arises, if only to be able to recognize the changes that inevitably occur. Decompression is thus the flip side of compression in that our need to achieve a holistic understanding of complex phenomena has sometimes to give in to the need to appreciate their inner complexity. (2012: 100)

Autobiographers have to ‘unpack’ their lives before the writing process begins, which usually requires them to take out the storage boxes and photo albums where the material anchors can be found. These may reveal details that were lost due to blending phenomena many years ago. Especially the self – the ‘I’ of autobiographical writing – has to be decompressed into various avatars and versions of the self that are likely to speak and act in sometimes radically different ways. The same applies to compromises or important decisions, which were preceded by conflicting points of view, which now have to be reconstructed to unravel their dramatic potential (cf. Dancygier 2005: 120). Writers may have a strong sense of the final blend, their present-day identities, but these have to be arrived at in a roundabout way: otherwise readers may consider these lives too streamlined and arranged (e.g. in the case of a hagiography), with all the elements carefully selected, sequenced and foregrounded in a strongly teleological manner. Yet, without a strong sense of identity, the autobiographical text would fall apart: “Compressing various images of ourselves along the dimensions of Time, Change, Cause-Effect, or Representation allows us to recognize the same person in a photograph of a five-year-old, in a valentine card written by a teenager to his sweetheart, and in a résumé attached to a job application” (2005: 102). It is indeed a difficult balancing act for autobiographers to find a meaningful compromise between compression and decompression and make their lives appear as a natural progression with emotional depth, despite the fact that they are heavily mediated and arranged.

To offer another example of compression from a semi-autobiographical text, I quote Kurt Vonnegut’s Slaughterhouse-Five, the second paragraph of chapter 1: “I really did go back to Dresden with Guggenheim money (God love it) in 1967. It looked a lot like Dayton, Ohio, more open spaces than Dayton has. There must be tons of human bone meal in the ground” (1991: 1). The last sentence is a great example of compression on a number of levels. The Dresden of WW II, of 1967 and the narrative’s present-day Dresden become blended, as do the three time periods. For the traumatised narrator 13 February 1945 exists in the present, which explains his observation that tons of human bone meal must ←218 | 219→be underneath his feet, which cannot be true in a literal sense. There is an immediacy and urgency to this statement that indicates that the narrator’s various war memories have become compressed into a single image of victimisation. Thus, narrative compression condenses experiences and feelings, it intensifies through selection and concentration. First-time readers may not know what to do with this statement or read it as some ominous foreshadowing. An intelligent way of building redundancy into a narrative is to work on the central themes at different levels of compression, not necessarily in the order of growing complexity. While the basic aim of reading is to integrate the particulars into more and more involved concepts and networks of meaning, blends triggered by literary texts may also require readers to unpack or decompress them. A successful blend maintains vital relations with its input spaces or, to put it in simpler terms, retains traces of its history. Like a modern-day city, the complexity of the present configuration reveals traces of the past and offers visitors different access points to explore its hidden structures.

Dancygier ascribes the narrator the role of the puppet master who is firmly in control of every aspect of the narrative: “The intentionality is crucial in that the very act of storytelling assumes the intention to use the story in its proposed form to communicate some content, even if various forms of narrative experimentation disrupt the impression of consistency and purpose” (2012: 59). She locates the narrator in an extradiegetic space that serves as a reference point for the various perspectives presented in the text: “The narrator is assigned to an independent space (story-viewpoint space; SV-space), which has the entire story in its scope. The story itself is contained in the main narrative space (MN-space), which consists of a number of narrative spaces (NS)” (2012: 38). Iser insists that the perspective of the narrator should not be confused with the meaning of the text (cf. 1980: 35), but Dancygier is very tempted to privilege the storyteller’s role and associates the final gestalt or ‘mega-blend’ with his or her perspective. Even if we replaced the narrator with the ‘implied author’, readers’ successful transactions with a text always transgresses the response a writer attempts to trigger through textual strategies.

Instead of narration, Dancygier’s approach is better suited to solve the problem of focalisation in autobiographical writing, where readers face a discrepancy between the narrator as the focaliser of the entire narrative and the various younger, experiencing selves with their own theories of the world and viewpoints. Her take on the coordination of perspectives seems more appropriate for this genre or traditional third-person novels than literature in general:

←219 | 220→

It is natural to assume that giving unmediated voice to characters constitutes some kind of stepping back on part of the narrator, a temporary yielding of the story viewpoint to a character. But there are reasons not to accept such an interpretation, since it would mean that characters’ discourse is beyond the scope of the SV-space, and, more important, that narrative fiction changes viewpoint levels all the time, without any mechanism ensuring the coherence which allows us to see how individual instances of direct speech are incorporated into higher narrative levels, including the MN-space. Such an explanation in general terms seems to be offered by the concept of viewpoint compression, as elaborated throughout this book. (2012: 141)

Since Dancygier understands reading as blending, she has to argue in favour of a process through which different viewpoints become compressed into a single vision (synopsis): “viewpoint compression is a blending mechanism which attempts to account for the fact that zillions of low-level facts, observations, or thoughts are compressed into more manageable viewpoint spaces and used in the processing of the narrative as a whole” (2012: 112). Since a translinear concept of meaning-making requires readers to recall and work with several narrative spaces at the same time, these have to be stored in compressed form to make them readily available.

The co-presence of external focalisation, which is tied to the narrator, and internal focalisation, such as a character’s perspective, is ubiquitous in the case of autobiographical comics, where the default setting is multiple and multimodal instances of focalisation: first-person verbal narration in text boxes, a (neutral) third-person visual point of view (cf. Mikkonen 2015: 106), a character’s direct speech and another character’s reaction, visualised through body language or verbalised in a speech or thought balloon, all within the same panel, is the norm rather than the exception. Kai Mikkonen speaks of “split verbal focalisation” (Mikkonen 2015: 103; see also Mikkonen 2008: 313) when the verbal narrator and a character express their points of view at the same time. In view of this complex layering of perspectives (cf. Mikkonen 2015: 106), Dancygier argues that we do not keep them separate throughout the narrative, but compress the viewpoints into a single meaning, which she ultimately associates with the narrator: “the multiplicity of viewpoints in narrative discourse is conceptually manageable because of a series of compressions bringing micro-level viewpoint up to the macro level of narrative spaces” (2012: 97; see also 67, 71–2). While a detailed narratological study of focalisation would introduce endless levels of complexity (cf. Zunshine 2006), readers easily manage to follow a scene, which means that they can keep track of whose points of view have been represented, but understand how the scene plays out as a unified whole. What is complicated to describe in terms of classical narratology, may be significantly easier to read.

←220 | 221→

Dancygier’s approach is clearly in line with enactivism and embodied cognition, which means that we learn to read and interact with others through direct encounters (cf. 2012: 112–6). She even states that the elaborate theories of Theory of Mind (ToM) are “an impediment rather than an aid to narrative comprehension” (2012: 112), as “the intersubjective (or ToM) skills develop in the process of understanding actions in context, and not in attempts to get into people’s heads” (2012: 114). She extends her criticism to the centrality of time in classical narratology and the meticulous reconstruction of the chronological order of events:

But in real situations we only follow the sequence to the degree to which we experience the events ourselves or are exposed to their results. Temporal sequence is rarely relevant to our understanding of events, but knowing their consequences is crucial. The sequence is a questionable criterion even in the most sequential of stories, but epistemic stance and understanding of causation seem to matter much more … (2012: 55)

By isolating the constituents of the story world and tracing their development across the entire narrative, narratologists tend to divorce them from their particular functions in specific configurations. Time can be an important element in the causal reconstruction of events, but otherwise readers are happy with the most basic and general orientation (cf. Emmott 2004: 47).

Marie-Laure Ryan, one of the leading transmedial narratologists, wrote an interesting essay for David Herman’s edited volume, Narrative Theory and the Cognitive Sciences, in which she discusses the importance of spatial orientation to readers’ experience of a narrative and whether it is true that readers build mental models of the environments described in fiction. She quotes Herman as one of the leading proponents of the idea that readers reconstruct the story world, including spatial configurations (Ryan 2003: 215), only to argue against this notion for the rest of the text. Choosing Gabriel García Márquez’s Chronicle of a Death Foretold as her main example, she starts her exploration with the following statement: “It takes a specific agenda – such as the present project – to attempt the systematic reconstruction of the ‘textually correct’ map of a fictional world. It was only on my third reading of Chronicle of a Death Foretold that I reached what I hope is a reasonably complete and accurate representation of the topography of the novel” (2003: 217–8). To come close to anything resembling the “model reader” or “super-reader” (2003: 218) of the novel, she had to work exceptionally hard and “perform several corrections” (2003: 218; see also 237). The problem with a detailed reconstruction of this particular map is the way the relevant information is scattered across the whole narrative (cf. 2003: 219–21) and directly impedes systematic spatial orientation. At the same time, Ryan felt the need to rise to the challenge and fill in the gaps: “The famed incompleteness ←221 | 222→of texts and the need to fill in informational gaps to reach a coherent interpretation is particularly acute when one tries to translate textual information into mental models of space, and these mental models into visual representations. A graphic map after all is not a cognitive map, but only the more or less faithful image of a cognitive map” (2003: 222). It is interesting that she assumes that a coherent interpretation has to involve topography, which, in this case, is obviously impossible for a highly trained reader like herself without marking all relevant data in the text, drawing numerous maps, cross-referencing the latest version with the book and repeating the process over and over again. Ryan asked a group of high school students to draw a map of the town after working on the novel for about three weeks and the results turned out exactly as one might expect: students developed very different map styles (e.g. tracing characters’ movements, sacrificing accuracy in favour of symbolic representations, adding illustrations or plot details), which demonstrates readers’ creativity and their unique approaches to literary texts more than anything else (cf. 2003: 228–30). Ryan draws the following conclusion: “The reader may thus be perfectly able to imagine the story’s main episodes without precisely situating each event on a global map” (2003: 235). The students were able to successfully and meaningfully discuss a novel of whose topography they only had a vague idea. This is possible, as Ryan suggests, when we acknowledge that readers are more concerned with the characters and their experiences than anything else in narrative fiction (cf. 2003: 236). While readers expect character arcs and trace them throughout the text, they are more reluctant to part with largely insufficient first impressions of a location: “I would like to speculate that once the map has been mentally sketched, it will be relatively resistant to new input or modifications. When new information conflicts with the reader’s mental model of space, it is easier to concentrate on the visualization of the current scene, and ignore the discrepancy, than to reorganize the whole map” (2003: 237). This serves as another reminder, much in line with Dancygier’s arguments, that the detailed narratological study of isolated elements is far removed from the experience of readers.

Since cognitive linguists claim that metaphorical thinking and conceptual integration are basic mental operations that do not necessarily require verbalisation, all of the theories presented so far also apply to other modes and multimodal narratives in equal measure:

The blend is characterized by its own structure (emergent structure), but can then become an input to another blend, or series of blends. While blending can account for ongoing processes of meaning construction (as in following the course of a conversation or reading a comic strip and accumulating content until the joke gels in the final frame), they can also become entrenched as new expressions, such as compounds or idioms. ←222 | 223→Blending also accounts naturally for multimodal contexts, where visual, linguistic, and aural prompts work in combination, as in film. (Dancygier 2012: 32)

As a consequence, narration as an analytical category becomes divorced from the idea that it has to be verbal. This is another important departure from classical narratology and aligns conceptual integration theory with transmedial narratology, where this is a basic premise. Here is Dancygier’s rationalisation:

… while language is naturally treated as the basic environment in which stories exist, it is not indispensable to narrativity, since a story can be ‘told’ through visual means. The concepts required for a narrative to emerge (sequentiality, causation, chunks of experience, cultural framing of such chunks, image-schematic force-gestalts of conflict and restored balance, et cetera) are the same concepts which are required for other language conceptualizations to emerge, so we should perhaps advocate a stronger claim, such that narrative form relies on a specific cluster of such concepts. (2012: 28)

In the context of aesthetic reading and reader-response criticism the only meaningful question is, of course, whether readers experience a text as a narrative or not. At the same time, the way that a narrative invites conceptual integration is equally relevant. While Dancygier associates the overall artistic unity, vision and viewpoint with the role of the narrator, that does not mean that he or she has to be verbally present all the time. Narrativity, she suggests, is bound to a cluster of ‘concepts’ that invite aesthetic reading rather than efferent reading, with a strong focus on character development (sequentiality, causation, experience, conflicts).

One of the more interesting observations in Dancygier’s study and the final point to be discussed here is ‘fictive vision’ (cf. 2012: 102–6). According to the logic of conceptual metaphor theory, we make sense of abstract or elusive concepts by ‘picturing’ them in more concrete terms, which can be expressed as understanding/knowing is seeing (e.g. I see what you mean.) (cf. Dancygier 2005: 106, 111). This aligns mental ‘perception’ (the mind’s eye), or rather conceptualisation, with visual perception. Based on a reading of her prose examples, Dancygier argues that narrators use that as a strategy and frequently describe imagined or counter-factual events, rationalisations, internal struggles etc. as ‘visions’: “There seems to be a consistent strategy at play here whereby mental representations and conceptualizations are narrated as perceptions” (2012: 102). Despite the ‘visualisation’ of concepts in prose texts, it is still the readers who have to picture the situation for themselves. In comics, however, the externalisation and materialisation of internal states, feelings and thoughts is the norm. What readers witness is not a more or less accurate depiction of the narrative’s story world, but direct access to how the narrator understands the feelings and thoughts of the characters, which is already a form of viewpoint compression, ←223 | 224→in the sense that the private thoughts of characters are mediated through the narrator’s visualisation of the narrative. They are narrativised in the sense that they have to fit into the narrator’s overall vision and are rendered in the narrative’s style. The most private thoughts – what in classical narratology would be internal focalisation – is very much externalised in comics, as experienced by the narrator:

The point of using mock-perception as a means of narrating conceptualization is not to give the reader access to a character’s thought processes, or to verbalized statements which count as thought-content, but to allow the reader to experience the narrated reality through the eyes of the narrating Ego. In other words, using visual construals as a means of conceptualization may give the reader the kind of insight which results not necessarily from access to thought processes, but from immediate access to experience. (2012: 103–4)

This also redefines the classical distinction between showing and telling. In visual narrative media, there may not be a verbal narrator, so everything is automatically a form of showing. However, the visualisation of internal states as external phenomena is very different from the idea of detached, camera-like or pseudo-objective narration that is usually associated with ‘showing’. I would argue that, in comics, the consistency in the depiction of characters, objects and locations is more indebted to readers’ need to be able to recognise them across panels than to the artists’ attempt to aspire to anything approaching realism. Even in so-called ‘non-fiction’ genres, such as autobiographical writing, documentaries, or reportage, there is often a strong tendency to sacrifice the markers of detachment and objectivity in favour of direct accessibility and the immediacy of experience – the ‘here and now’ vs. the ‘there and then’, which can generally be understood as a tension between the past tense of the narrator’s voice and the immediate presence of the depicted scene. In this sense, narratives present their central issues or themes by dramatising them, e.g. staging problems or disputes as confrontations between characters. While this is blatantly obvious in drama, Dancygier describes this as a general characteristic of fiction:

It is not at all surprising that narrative choices would capitalize on the links between experience and higher cognition by appealing to the reader’s experiential abilities, rather than rely on the ‘telling’ technique. It is thus possible to redefine and expand the idea of ‘showing’ by applying it to narrative choices which prompt experiential alignment. (2012: 106)

It is not a coincidence that Fauconnier and Turner, but also Dancygier choose the performing arts to illustrate the core principles of blending theory. The stage is a prime example of a blended space where, in the imagination of the audience, ←224 | 225→actors become characters: “Dramatic performances are deliberate blends of a living person with an identity. They give us a living person in one input and a different living person, an actor, in another. The person on stage is a blend of these two” (Fauconnier & Turner 2003: 266). For the duration of the play audiences embrace the idea that what is unfolding in front of their eyes is real enough for them to be affected by the developing narrative, which means that “the ability to live in the blend provides the motive for the entire activity” (2003: 267). It requires spectators’ active participation, in the sense that “the drama story is a blend emerging out of the audience’s interaction with on-stage prompts and offstage frames” (2012: 145). Thus, characters and props function as material anchors (cf. 2012: 147, 158), especially when the latter are foregrounded by the stage action and then reappear in a later scene, so that they connect important elements of the play with each other and make theatre-goers recall important contexts from previous scenes (cf. 2012: 157). Live performances are also a great illustration of Dancygier’s concept of ‘showing’: “The visual choices made by the director help underscore the claims made above: the discourse of the play needs to represent the hidden inner thoughts of characters, and often relies on material objects to serve as addressees in linguistic constructions specializing in thought representation” (2012: 161). The soliloquy is just one of many strategies to externalise and visualise – through performance – the inner thoughts and feelings of characters and make them accessible to the audience.

Material aspects of theatrical space are exploited to profile subjective construals beyond the characters’ words and play a central role in prompting story construction processes. The reliance on materiality has a clear advantage over relying on discourse alone, since supporting psychological components of the story with speech and objects conveys information to the audience through multiple channels. Also, stage time is allocated to events based on their emotional impact, not verisimilitude. (2012: 164)

This last point is crucially important: in the theatre, showing has nothing to do with realism. Live performances are as artfully constructed as all narratives, but “the holistic nature of the theatrical experience” (2012: 142) allows for a vicarious entanglement in the ongoing action that is far removed from narratological analysis. A basic tenet of Dancygier’s theory and reader-response criticism in general, for that matter, is that this insight applies to all narratives, even if it is more evident in a live performance or in the cinema, where audiences cannot stop the narrative progression to second-guess or deconstruct the presented action.

As a bridge to the next part, I briefly introduce Amy Spaulding’s The Page as a Stage Set: Storyboard Picture Books, in which she explains dramatic changes – pun ←225 | 226→intended – in picture books as a break with traditional storytelling in the late 1970s and early 80s:

When this [an increase in dialogues] joins the fact that the plot is being acted out as much as recounted by a narrator, it becomes evident that picture books are growing farther away from illustrated novels and closer to drama.

There is a type of picture book in which this dramatic leaning is very noticeable: those books which have adapted elements of the comic strip. The adaptation of comic book elements in these books is particularly interesting in relation to the idea, already mentioned, that a form of picture book is evolving in which the pictures and words join together to tell the story. Like picture books, the comics use pictures and text to tell their story, but they have grown out of the tradition of caricature and cartoon, rather than out of the novel and folktale as picture books did, so the emphasis on the art as the primary giver of information is natural.

Picture books and comics together represent a unique means of telling a story, a form of print narrative that is in many ways more dramatic than literary. Although they physically appear as books, they imitate dramatic performance, and the drawings become the counterparts of what would be presented in the staging of a theater or film production. (1995: 5)

Spaulding resorts to the terminology of film studies and calls these hybrid forms ‘storyboard books’, thus conceptually bridging the gulf between picture books, comics, film and theatre performances. What ties these visual narrative media together is an increasing reliance on showing: as Dancygier argues, even in prose a strong tendency towards visualisation and ‘acting out’ scenes can be detected, but this becomes much stronger in the context of visual narrative media, ranging from a balance in traditional picture books to pure performance in staged plays. This is how Spaulding comments on the transformation: “Traditional narrative forms tend to give the audience a sense of detachment, of reading something which has already concluded, while the dramatic forms gives [sic] a sense of immediacy, of watching a situation unfold as it is is [sic] happening” (1995: 15). Generally speaking, comics take the exact middle ground on this scale, wedding traditional narration with dramatisation – literally and figuratively – but more often than not leaning towards the latter. Readers who encounter comics for the first time tend to treat them as heavily illustrated prose texts and forget that they are also stage plays in a sense. The comics historian Robert Petersen, for example, believes that the “rise of the graphic narrative in Europe parallels the rise in popular melodrama, and it is in this direction that we find some important similarities” (2011: xvi). He goes on to argue that “graphic narratives rely on representing things in a way that is predicated on our cognition of how we make sense of our known world. In this respect, the visual elements in a graphic narrative are like objects on stage: they are animated with potential ←226 | 227→signification, adopting meanings beyond what they may simply represent in the everyday world” (2011: xvii). Writing about the hybridisation of picture books and comics, Spaulding offers the following observation: “Staging/Design is very important to any dramatic production, since it provides mood and flavor; in picture books the staging is an important part of the basic narrative. The author/artist, as playwright, also becomes director and cameraman, set designer and casting director, location manager and even costumer” (1995: 37). In the following part we look at how cognitive theories and stage metaphors help to make sense of a medium that often combines retrospective verbal narration with the immediacy of dramatisation.

←227 | 228→