Stylistics

6 Text and cognition II

Text processing

6.1 The reading process

In the previous chapter we considered the active role that readers play in the construction of meaning. We focused on the prior knowledge that readers bring to texts and which they use in the process of interpretation, and from this it becomes clear that the process of meaning creation is a result of the interconnection between textual triggers and readers’ world knowledge. Or, to restate this in Semino's (1997) terms, texts project meaning while readers construct it. The means by which readers go about constructing meaning is, as we explained in Chapter 5, the central concern of cognitive stylistics. In this chapter we will continue our consideration of this branch of stylistics by focusing on how readers navigate their way though texts. While Chapter 5 considered the stylistic effects that can arise as a result of, say, deviant schemas or novel conceptual metaphors, in this chapter we will focus primarily on a descriptive account of how readers process textual meaning. In so doing we will outline some of the most influential theories of text processing to have been adopted by stylistics. One caveat to the whole cognitive stylistics enterprise, of course, is that it is important that it does not reject the more linguistically and textually orientated approaches described in the earlier chapters of this book. Rather it should seek to enrich these by adding a cognitive layer to the explanation of how readers react to texts.

6.2 Text world theory

Text world theory is a theory of discourse processing developed initially by Paul Werth in a series of papers (see, for example, Werth 1994, 1995, 1997) and outlined in full in his posthumously published book, Text Worlds: Representing Conceptual Space in Discourse (Werth 1999). Werth's aim in the development of text world theory was to account for how we as readers and hearers ‘make sense of complex utterances when we receive them’ (Werth 1999: 7). His suggested answer was that we do this by constructing ‘mental constructs called text worlds’ (1999: 7). Werth defined text worlds as ‘conceptual scenarios containing just enough information to make sense of the particular utterance they correspond to’ (1999: 7). The cognitive nature of the theory is made clear by Werth in his clarification of the necessity of three elements – author, text and reader ¹ – in the creation of a text world:

the author creates only a text; he/she will have a particular text world in mind, but there is no guarantee at all that the reader will manage to produce the same text world on reading that text. We cannot say that the author's text world is the definitive one, since, in fact, there is no such thing. We may say, therefore, that a text world does not come into being until each of the three elements – author, text and reader – are present.
(Werth 2007: 155)

Werth's work has been particularly influential within stylistics and has been taken up enthusiastically by a number of researchers (see, for instance, Gavins's 2007 subsequent development of text world theory). At its heart is the conceptual metaphor THE TEXT IS A WORLD, and text world theory is an exposition of how readers (and hearers) mentally construct such worlds as a means of interpreting discourse in context.

Text world theory makes an initial distinction between the discourse world and the text world. The discourse world is the immediate real-world situation in which a writer communicates with a reader. The text world is what is constructed by the reader to make sense of the communicative event. Included within the discourse world is the schematic knowledge of all participants in the discourse, as well as all surrounding physical objects and entities, and together these form a context. Werth defines context as ‘the relevant situational background(s) for and in a particular discourse’ (1999: 117). The key word here is ‘relevant’, since the potential context for any given discourse world is vast. Discourse participants restrict this by only considering common ground information: i.e. only that information which is necessary for the interpretation of the discourse in question.

Participants in the discourse world use the textual and common ground information present within it to construct a text world – a mental representation of the text. Text worlds are composed of world-building elements and function-advancing propositions, both of which are recovered from the text. World-building elements consist of time (realised through the tense and aspect of verb phrases), location (realised through adverbials and noun phrases specifying place), characters (realised through proper nouns and pronouns) and objects (realised through nouns and pronouns). Function-advancing propositions work to develop and advance events within the text world, and are realised in verb phrases. Here it is useful to return to the functional grammatical categories introduced in our discussion of transitivity in Chapter 3, since function-advancing propositions can be seen as mapping onto these Hallidayan processes. Function-advancing propositions may take the form of material processes (that is, intentional, superventional or event processes), relational processes (intensive, possessive and circumstantial) and mental processes. In text world theory, material processes are indicated diagrammatically by vertical arrows, while relational and mental processes are indicated by horizontal arrows. An example of a text world can be seen in figure 6.1, which represents diagrammatically the opening paragraph of Thomas Hardy's The Mayor of Casterbridge:

(1)

One evening of late summer, before the nineteenth century had reached one third of its span, a young man and woman, the latter carrying a child, were approaching the large village of Weydon-Priors, in Upper Wessex, on foot. They were plainly but not ill clad, though the thick hoar of dust which had accumulated on their shoes and garments from an obviously long journey lent a disadvantageous shabbiness to their appearance just now.
(Hardy 1987[1886]: 1)

Figure 6.1 Text world (including temporal world-switch) generated through reading the opening paragraph of The Mayor of Casterbridge

What is apparent in figure 6.1 is that processing this opening paragraph involves the setting up of two text worlds. The first (text world 1) is generated by the third-person past-simple narration and includes the characters of the young man and woman and the child, as well as the specification of location and time, the characters’ shoes and garments and the ‘thick hoar of dust’. However, the past perfect verb phrase ‘had accumulated’ requires the reader to generate a second text world (text world 2), wherein the dust accumulates on the shoes of the young man and woman. Gavins (2007) notes that a change in tense such as that in example 1 invokes a temporal world-switch that results in the instantiation of a further text world. The world-switch is indicated in the diagram by a rounded rectangle. Gavins introduces the term world-switch to replace Werth's preferred sub-world on the basis that subsequently developed text worlds are not necessarily subordinate to the primary text world. Following this logic, we use Gavins's term here. As Gavins (2007: 52) explains, world-switch also maps onto Emmott's (1997: 147) notion of a frame-switch in contextual frame theory (see section 6.4), thereby providing a useful link between these two approaches to text processing.

Example 1 constitutes a relatively simple text in processing terms, though text world theory is also able to account for much more complex discourse through its concept of modal worlds. Drawing on Simpson's (1993) modal grammar and adapting Werth's (1999) notion of sub-worlds, Gavins (2007) postulates the concepts of boulomaic, deontic and epistemic modal worlds. Respectively, these modal worlds represent the desires, obligations and beliefs of participants in the discourse world or characters in the text world (see Chapter 3 for full definitions of boulomaic, deontic and epistemic modality). These modal worlds are generated as we read example 2, an extract from Larry McMurtry's novel Lonesome Dove:

(2)

[Lonesome Dove is set in the American West of the mid-1800s. Newt is a young cowboy with the Hat Creek Cattle Company outside the town of Lonesome Dove. Lorena is a prostitute with whom he is infatuated.]
On an average day, Lorena occupied Newt's thoughts about eight hours, no matter what tasks occupied his hands. Though normally an open young man, quick to talk about his problems – to Pea Eye and Deets, at least – he had never so much as uttered Lorena's name aloud. He knew that if he did utter it a terrible amount of ribbing would ensue, and while he didn't mind being ribbed about most things, his feeling for Lorena was too serious to admit frivolity. The men who made up the Hat Creek outfit were not great respecters of feeling, particularly not tender feeling.
There was also the danger that someone might slight her honor. It wouldn't be the Captain, who was not prone to jesting about women, or even to mentioning them. But the thought of the complications that might arise from an insult to Lorena had left Newt closely acquainted with the mental perils of love long before he had had an opportunity to sample any of its pleasures except the infinite pleasure of contemplation.
Of course, Newt knew that Lorena was a whore. It was an awkward fact, but it didn't lessen his feeling for her one whit.
(McMurtry 1985: 24)

There are a number of world-building elements in this passage. Time and location are recoverable from what we have read previously and so we know that the story as a whole takes place in a town called Lonesome Dove somewhere in the American West in the mid-nineteenth century. Time is specified more specifically through the adverbial ‘on an average day’. What this also suggests is that we are not dealing with what we might term a concrete text world. The text world presented in this paragraph is a compression of a number of other text worlds in which the action of Newt daydreaming about Lorena has happened. With regard to characters, the primary ones in this text world are clearly Newt and Lorena, though others are mentioned (Pea Eye and Deets, the Captain and ‘The men who made up the Hat Creek outfit’). The mention of these other characters also helps to specify place, since these characters are anchored to a specific location. Newt is a cowboy with the Hat Creek Cattle Company, as are Pea Eye, Deets and the Captain, therefore mention of Pea Eye, Deets, the Captain and the Hat Creek outfit provides a specification of location as a result of our pragmatic knowledge of these characters.

In addition to the world-building elements that help us to populate this text world, the modal verbs in the passage trigger the generation of a number of modal worlds, all of which are as rich as the primary text world of which they are offshoots, in the sense that they too contain world-building elements and function-advancing propositions. For example, the sentence ‘There was also the danger that someone might slight her honor’ triggers a boulomaic modal in which another member of the Hat Creek outfit insults Lorena and Newt is forced to acknowledge his love of her by defending her honour. ‘It wouldn't be the Captain’ triggers a negative epistemic modal world in which the Captain does insult Lorena (Chapter 3, section 3.2, considers more fully the local textual functions of negation; see also Nørgaard 2007 and Nahajec 2009). In causing us to generate such a modal world, the text foregrounds the unlikelihood of this event ever happening, thereby giving us an indication of the Captain's character. A further epistemic modal world is triggered by the sentence ‘Of course, Newt knew that Lorena was a whore.’ This reference to what Newt believes to be the case about the world that he inhabits generates an element of conflict in the story. Newt would like nothing more than to marry Lorena, though the content of this epistemic modal world conflicts with the content of the boulomaic modal world in which he and Lorena live happily ever after.

What is also apparent in this passage is that the modal worlds generated are not participant accessible. That is, the reader has no way of verifying the truth of these modal worlds, since the content of these is inaccessible to participants in the discourse world. Rather, these are character accessible only, hence Newt's beliefs about what would result from a declaration of his love for Lorena are necessarily true only in his modal worlds. We have no means of determining whether this would be the case in the framing text world, were it to happen.

The generation of a number of modal worlds in this passage has several effects. It indicates, for example, that Newt is a character who is thoughtful and not prone to impulsive behaviour. It also generates some of the conflict that is necessary for the advancement of the narrative. Of course, as Stockwell (2002a: 145) notes, a text world theory analysis is not necessarily required to uncover this kind of analytical detail. Nonetheless, text world theory does offer a principled explanation of the way in which we might keep track of narrative information as we read. One of the key aspects of text world theory is the concept of the reader toggling between different text worlds, modalised or not, in order to process the events that constitute the narrative. This will result in several shifts of position within the text world in order to take account of the viewpoints of a variety of characters (note, too, that this is also the case for non-fiction texts). For this reason, we will focus in the next section on a theory of text processing that aims to account more specifically for the means by which readers shift mental positions as they read.

6.3 Deictic shift theory

Deictic shift theory was originally developed by an interdisciplinary research group consisting of linguists, psychologists, computer scientists and literary critics, among a number of others. The theory as a whole is outlined in a collection of papers edited by Duchan et al. (1995), though for reasons of space we concentrate here specifically on those aspects of the theory that have had applications with stylistics. What we will also attempt to do is demonstrate how deictic shift theory might be used to add a cognitive layer of explanation to how readers process point of view when they read.

Deictic shift theory was developed to account for the way in which readers can come to feel deeply involved in what they are reading, to the extent that they forget about their position within the real world (the discourse world, in text world theory terms) and begin to interpret events in the narrative as if from a position within the text world. If text world theory provides an explanation of the means by which readers make sense of a narrative, then deictic shift theory provides a means of explaining the sense of involvement that readers often have as they navigate their way through that narrative. Segal, one of the original proponents of deictic shift theory, suggests that the reader ‘takes a cognitive stance within the world of a narrative and interprets the text from that perspective’ (Segal 1995: 15). At the heart of deictic shift theory is, as the name suggests, the concept of deixis.

Prototypically, deixis refers to the linguistic encoding of spatial and temporal relations between objects and entities. Deictic terms are the specific linguistic expressions that encode this information. Following Levinson (1983), we can distinguish five different types of deixis: (i) place deixis, (ii) temporal deixis, (iii) person deixis, (iv) social deixis and (v) empathetic deixis. The first of these encodes the relative position in space of specific objects or entities. Pure place deictics include adverbs such as here and there which can only be interpreted by reference to the position of the speaker or writer of these words. This and that are further examples of pure place deictics, since they rely for their interpretation on a knowledge of the location of the speaker or writer. Pure deictics tend to come in pairs, one of which will indicate proximity to the speaker or writer in question, while the second indicates distance from that speaker or writer. Locational deictic expressions are slightly different in nature in that they are interpreted in relation not to the speaker or writer but to the position of other referents within the situational context. Consider, for example, the adverbials in example 3:

(3)

She had a sign painted in bright colours, which was then set up just off the Lobatse Road, on the edge of town, pointing to the small building she had purchased: THE NO. 1 LADIES’ DETECTIVE AGENCY.
(McCall Smith 1998: 4)

The prepositional phrases ‘off the Lobatse Road’ and ‘on the edge of town’ can only be interpreted if we know where both the Lobatse Road and the town are situated within the text world. Since this is information included in the world-building elements of the text world, these expressions are not difficult to interpret. The point is that they are not reliant on our knowing the position in space of the narrator, as expressions such as here and there are. For this reason, they are not considered to be pure deictics.

While place deictics encode physical position within the situational context, temporal deictics encode distance in time from the moment of speaking or writing. Pure temporal deictics include now and then, and yesterday, today and tomorrow. Temporal deixis can also be expressed through such time expressions as, for example, a week ago, some years previously, etc. In the same way that temporal deixis encodes metaphorical distance from the moment of speaking, person and social deixis encode the speakers and hearers within a situational context and their relative distance from the speaker or writer. Person deixis is realised through personal pronouns (e.g. I, you, he, she, etc.). Social deixis is related to person deixis and encodes perceived social distance between particular characters.² This may be realised through, for example, the use of titles and honorifics to encode social distance (e.g. Professor Jeffries, Dr McIntyre) and the use of first names and short forms to indicate social proximity (cf. Lesley and Dan). Finally, empathetic deixis encodes psychological attitude. Compare, for example, the difference between ‘So, tell me about this new colleague of yours’ and ‘So, tell me about that new colleague of yours’. The second may generate a sense that the speaker harbours less positive feelings towards the colleague in question than the first (though context too may play a part in the choice of demonstrative here).

Pure deictics, as we have explained, are interpreted in relation to the position in space and time of the speaker or writer. This position is known as the deictic centre and is a central concept in deictic shift theory. It is our knowledge of the deictic centre to which deictic expressions refer that allows us to interpret them. In everyday conversation we have a tendency to regard ourselves as the deictic centre for all deictic terms. However, we also have the capacity to suspend our egocentric conception of deixis and project a deictic centre that is not our own; that is, we are able to interpret deictic terms as having a reference point other than that which we might prototypically expect. As an example of this, imagine receiving instructions over the phone from an IT technician about how to solve a problem with your computer. Such instructions might include phrases like ‘Remove the flash drive from the USB port on your left and try the USB port on your right’. In such an instance, the person giving the instructions would be using locative deictic expressions relevant to your deictic centre as opposed to their own. In effect, they would be engaging in deictic projection. This capacity for deictic projection is at the heart of deictic shift theory, since the theory suggests that readers are able to feel involved in a narrative by experiencing vicariously events from a viewpoint other than their own. Clearly, this involves deictic projection.

Deictic shift theory attempts to spell out precisely how readers understand projected deictic centres. In her explanation of how this happens, Galbraith (1995) suggests that narratives have numerous deictic fields, which Stockwell (2002a: 47) defines as a set of deictic expressions all relating to the same deictic centre. The term deictic field comes originally from the work of Bühler (1982 and 1990 [1934]). In fiction deictic fields are related either to characters, narrators or narratees, and in non-fiction to what text world theory would term enactors – essentially, non-fictional characters within the text world in question. In Galbraith's (1995) summary of Bühler's work she notes that there are three aspects to a speaker's deictic field. The first is the immediate physical space and the immediate time in which a speaker is located. This aspect of the deictic field is realised linguistically through the use of pure deictic terms such as here and now. The second aspect is the pragmatic context in which a speaker is communicating. Galbraith notes that an awareness of this pragmatic context enables a speaker's addressees to understand the speaker's deictic references. Finally, the third aspect of any given deictic field is its capacity to extend beyond a speaker's immediate physical and temporal location, and to refer to that speaker's memory or imagination. This is a much simplified summary of Bühler's definition of a deictic field, though it will suffice for our purposes (for a full discussion and critique of the notion of a deictic field see McIntyre 2006: 99–107).

Deictic shift theory suggests that readers assume the spatial, temporal and social coordinates of deictic fields not to be egocentric and related to the discourse world, but to be anchored within the text world itself. Suspending our normal assumptions about deictic reference involves making a deictic shift into the text world, in order to take up the cognitive stance that Segal (1995) explains results in a feeling of involvement in the narrative. McIntyre (2006: 107) notes that while it is not possible in real life to inhabit exactly the same deictic field as another person, in fiction this limitation does not apply and it is entirely possible to take up a position within a character or enactor's deictic field as we read. It follows that deictic shifts must be triggered somehow and that such triggers may be linguistic or non-linguistic in type. The first deictic shift we make when reading, for example, is triggered simply by picking up a book and starting to read. Once centred within the text world, we move around it by shifting into and out of the deictic fields of which the narrative is composed. And if whatever deictic field we happen to be in is not regularly reinstantiated by further deictic reference, it will eventual decay. This process can result in interesting stylistic effects and is particularly prominent as a technique in many contemporary films. For example, Roberto Benigni's Life is Beautiful begins with voiceover narration. We therefore understand that the fictional world we see on screen is embedded in the framing narration of the voiceover. However, once the action begins, the voiceover is not continued. The deictic field of the framing narrator thus decays, and we forget that we are watching an embedded narrative. Only at the end of the film is the voiceover narration reinstantiated, causing us to shift deictic fields again, from the embedded narrative level to the framing narrative level.

The means by which we shift around the deictic fields of a text are described by the terms PUSH and POP. A PUSH is defined by Galbraith as a movement ‘from a basic level to a less available deictic plane, such as episodic memory (known as “flashback” in fiction), fictional story world (this may be a fiction within the fiction), or fantasy’ (Galbraith 1995: 47). Conversely, the term POP refers to movement out of a particular deictic field. According to deictic shift theory, we PUSH into deictic fields and POP out of them. We can see how this works if we consider an example:

(4)

During the night of the fifth of December, the train ran in a south-easterly direction for about fifty miles; then, going up a like distance to the north-east, it drew near to Great Salt Lake.
About nine o’clock in the morning, Passepartout went out on the platform to take the air. The weather was cold, the sky grey, but it was snowing no longer. The sun's disc, magnified by the mist, looked like an enormous gold coin, and Passepartout was busy calculating its value in pounds sterling, when his attention was diverted from this useful occupation by the appearance of a somewhat strange personage.
This man, who had taken the train for Elko station, was tall, very dark, with black moustache, black stockings, a black silk hat, a black waistcoat, black trousers, a white tie and dog-skin gloves.
(Verne 1994 [1873]: 110–11)

This extract from the beginning of Chapter 27 of Around the World in Eighty Days is a third-person narrative in which Passepartout's point of view of the events in the text world is prominent. As a reader, we are likely to feel fairly involved in the text world because of our capacity for interpreting the deictic expressions within it as referring to the character of Passepartout's deictic centre. The adverbials ‘During the night of the fifth of December’ and ‘About nine o’clock in the morning’ specify the temporal coordinates of Passepartout's deictic field, and our knowledge of the story so far allows us to interpret these more fully as occurring in the year 1872. The other deictic terms in the passage also relate to Passepartout's deictic centre, meaning that to interpret these we have to take up a cognitive stance in the text world that is the equivalent of Passepartout's. For example, the narrator tells us that ‘Passepartout went out on the platform’. The place-deictic verb went indicates movement away from a deictic centre and generates the sense that we the reader are in the same deictic position as Passepartout. As a means of further verifying this, consider the difference in effect had the sentence been ‘Passepartout came out on the platform’. This sentence gives the impression that we are already positioned on the platform and Passepartout is moving towards us.

The more deictic information we are provided with by a text, the richer our conception of the deictic field will be, thereby increasing the degree to which we feel involved in the text world. For instance, the proximal place deictic this is used to refer to the man that Passepartout notices, suggesting his physical proximity to Passepartout. Again, to compare an alternative rendering of this, consider the effect generated had the sentence been ‘The man, who had taken the train for Elko station [. . .]’. The deictic references in the text, then, assist the reader in constructing a relatively rich deictic field, at the deictic centre of which is Passepartout. To use the terms of deictic shift theory, we PUSH into Passepartout's deictic field. The cognitive stance that we take up within the text world is reinforced by the other point-of-view indicators in the passage that reflect Passepartout's viewpoint. For example, Passepartout's first sight of the man in black is described using indefinite reference (‘a somewhat strange personage’). Indefinite reference is typically used to introduce new information, and because the sight of the man constitutes new information for Passepartout, it is presented to the reader as new information too in order to reflect Passepartout's point of view.

At this point we are in a position to draw some parallels between deictic shift theory and text world theory. Text world theory provides a principled account of how we conceive of the meanings projected by texts. Conceiving of a text world involves taking up a cognitive position within it. Deictic shift theory specifies how we do this through its concept of deictic fields and the notion of pushing and popping between these. There is, though, a problem with the notion of POPs and PUSHes. McIntyre (2006) discusses this in detail and we summarise the issues here.

The first problem is that there are different types of PUSHes and POPs. So, for example, a flashback in a narrative constitutes a PUSH back along the temporal continuum of a story. However, a move out of the fantasy world of a character and back into the basis-level text world constitutes a POP up a level of discourse. At the very least, then, deictic shift theory needs to specify the kinds of PUSH and POP that readers engage in.

A more serious problem concerns the nature of these cognitive movements as a whole. One unfortunate consequence of the terms PUSH and POP (both of which were borrowed from computer science) is that while PUSH suggests an active movement on the part of the reader, POP suggests a movement that is involuntary. Cognitive stylistics puts substantial emphasis on the active role of the reader in text processing and so the notion that a cognitive movement may be involuntary seems counter-intuitive. Furthermore, it must be the case that such movements occur as a result of triggers, be these linguistic or non-linguistic. It would seem, then, that there is little difference in theoretical terms between a move into a deictic field and a move out of one. Both are cued by a trigger and both are active movements on the part of the reader. For example, a POP out of one deictic field must involve a PUSH into another. We suggest, then, that it may be more useful to disregard the terms PUSH and POP and refer instead simply to deictic shifts, specifying the type and direction of these (see McIntyre 2006: 108–12 for more detail on this issue).

6.4 Contextual frame theory

The final theory of text processing that we will consider in this chapter is Emmott's contextual frame theory (outlined in full in Emmott 1997). We should point out, of course, that none of the theories discussed in this chapter exist in isolation from each other. Text world theory, for example, is designed to explain how we construct rich mental models as we read, while deictic shift theory complements this by providing an explanation for why we can come to feel deeply involved in the text worlds of a narrative. By the same token, contextual frame theory provides a further connection between these theories by providing an explanation of how we keep track of the elements of a narrative as we read. Of course, the complexities of such an activity are vast and it would be reasonable to ask what contribution stylistics might make to our understanding of this process, especially when the insights provided by psychology into the nature of reading remain at a relatively basic level. Emmott herself suggests an answer when she explains that ‘[w]hilst textual analysis cannot prove what goes on in a reader's mind, it can reveal the complexity of texts and can thereby indicate the nature of the task to be undertaken’ (Emmott 1997: 99). In this respect, the insights of cognitive stylistics might be used to form hypotheses about the nature of reading that would then be available for testing by empirical psychologists and linguists.

Contextual frame theory suggests that readers make sense of narrative events by relating them to the contexts in which they occur in the text world. To do this, readers construct contextual frames. Emmott (1997: 121) describes a contextual frame as ‘a mental store of information about the current context, built up from the text itself and from inferences made from the text’. Contextual frames contain information which may be episodic or non-episodic in nature. Episodic information is that which is true at a particular point in the narrative in question, but which is not necessarily relevant beyond this point. Non-episodic information, conversely, is that which is true ‘beyond the immediate context’ (Emmott 1997: 122). As an example, in Jonathan Coe's novel The Rotters’ Club, the fact that the main protagonist, Benjamin Trotter, is a schoolboy growing up in the 1970s constitutes non-episodic information: i.e. it is information which remains true across the course of the narrative as a whole. We may compare this with an instance of episodic information. In Chapter 7 of the novel, Benjamin arrives for a school swimming lesson only to find that he has forgotten his swimming trunks. Since the punishment for this is to swim naked, Benjamin is mortified and prays to God for an answer to his dilemma. As he does so, a locker door swings open and Benjamin sees that inside the locker is a pair of swimming trunks. This, then, is episodic information: i.e. a one-off occurrence in the narrative. While Emmott notes that episodic information does not have to have relevance beyond its immediate context, in this particular case it does, since the discovery of the swimming trunks leads Benjamin to believe that his prayer has been answered. As a result of this, Benjamin becomes convinced of the existence of God and embraces Christianity, an action that is to have a significant impact on his behaviour throughout the remainder of the novel.

In addition to the episodic and non-episodic information stored within contextual frames, Emmott explains that such frames also hold information relating to the situation in question as well as details of descriptions and events recently referred to. Such information is necessary for a reader to interpret anaphoric references, as may be seen in the following example:

(5)

By everyday standards, it was an exceptionally disastrous swimming period. Benjamin was chosen to be part of a relay team captained by Culpepper, and was clearly the weakest link in the chain. By the time he had completed his two purple-faced, asphyxiated, floundering lengths of butterfly stroke, their lead had been all but erased[.]
(Coe 2004: 72)

In just this short extract we can see that the pronoun he in the third sentence relies for its interpretation on the reader having retained from the preceding sentence the frame knowledge of Benjamin being part of a relay team (there is a clear link between this cognitive concept and the linguistic notion of cohesion; see section 3.2.3). This enables us to link he with both the proper noun Benjamin and the noun phrase the weakest link in the chain. Note too that the description in the preceding sentence of Benjamin as a weak swimmer allows us to infer that he refers to Benjamin as opposed to Culpepper, since this description is consistent with the negatively-charged adjectives purple-faced, asphyxiated and floundering used to describe Benjamin's swim in the latter sentence.

Emmott explains that this kind of contextual monitoring allows readers to create an overall context for the narrative. However, while it is possible for readers to hold information about more than one context at any given time, we tend to focus our attention on one particular context. The terms binding and priming explain the means by which we do this. Binding refers to the process of linking particular entities to particular contexts, while priming refers to the process by which particular contextual frames are brought to the forefront of the reader's mind. For example, Ian Fleming's James Bond novel, Dr No, begins with a scene in a Jamaican street in which three blind beggars make their way slowly along the road. These three characters are bound into this particular situational context. The scene then switches to a gentlemen's club where Brigadier Bill Templar, John Strangways, a professor from Kingston University and a criminal lawyer are playing cards. These four characters are bound into this second situational context. At this point in the reading process, the scene in the cardroom of the gentlemen's club constitutes the primed context, since this is the context on which the reader's attention is currently focused. The characters of Strangways, Templar and the two others are therefore primed as well as bound. Nonetheless, the three beggars of the first scene remain bound into their own particular context, despite the fact that it is now unprimed.

Binding and priming are thus processes of monitoring that readers engage in. Characters that are primed will remain in a reader's mind even if they are not referred to directly in the text. A primed character that is referred to in a sentence is said to be overt. One that is present in the scene but not referred to is said to be covert. If we return to our Dr No example, in the following paragraph Templar is the overt primed character, since he is referred to directly in the proper noun and its appositive noun phrase, and by the pronoun he:

(6)

Bill Templar, the Brigadier, laughed shortly. He pinged the bell by his side and raked the cards towards him. He said, ‘Hurry up, blast you. You always let the cards go cold just as your partner's in the money.’
(Fleming 1985 [1958]: 7)

The other characters (Strangways, the professor and the lawyer) are, at this point in the narrative, covert, since they are not mentioned in the text but remain primed; that is, we remain aware of their presence within the scene, despite the fact that the text does not refer to them. Only primed characters may be overt or covert. Unprimed characters – that is, characters not present in the primed frame but bound into some other context – are neither overt nor covert, but simply bound into the reader's main context.

Contexts such as those described above are, of course, continually modified across the course of a whole narrative. The reader's attention will shift to new contexts, and these new contexts will involve new entities and characters. The means by which readers track such context changes is also covered by contextual frame theory. Emmott suggests, for instance, that frame modification occurs when characters enter or leave a particular location. For example, the sentence that follows the one in example 6 describes Strangways leaving the club: ‘Strangways was already out of the door.’ As a result of this, the reader modifies the contextual frame in which Strangways was previously bound. The other elements of the frame remain in place – we assume, for instance, that Templar, the professor and the lawyer remain in the cardroom – and so the frame is not radically altered but simply modified by Strangways becoming unbound from the frame.

This example is an instance of the modification of a primed frame, though unprimed frames may also be modified. However, this cannot be done via overt reference or it would prime the frame in question. The way in which unprimed frames can be modified without consequently priming them is via covert reference. For example, in Chapter 2 of Dr No the character of James Bond enters the narrative. Bond, at this point, is in London, hence he is bound into this contextual frame. At the beginning of Chapter 4, however, Bond is on board a plane to Jamaica. We assume, therefore, that he has become unbound from the previous contextual frame (the London scene), despite the fact that there was no overt reference to this happening. (Note that frame modification is similar to the notion of tuning a schema. Some variants of schema theory even make use of the term frame to describe schematic knowledge; see section 5.2.1 for details.)

Frame modification, then, refers to the adding or removing of a character from a particular frame. The notion of a frame-switch refers to the process by which a reader stops monitoring one frame and starts monitoring another. Frame-switches generate new frames, usually leaving the previously established frames intact (albeit potentially modified). Frame-switches may be instantiated pragmatically or via adverbials of time or place. They may also be instantaneous or progressive. Instantaneous frame-switches involve a sudden mental leap in space or time, such as that triggered by the mention of Bond being on a plane to Jamaica rather than in an office in London. A progressive frame-switch is one which does not involve such a mental leap, but where the temporal or locative change is tracked explicitly, as in the following example:

(7)

Strangways [. . .] walked quickly across the mahogany panelled hallway of Queen's Club and pushed through the light mosquito-wired doors and ran down the three steps to the path.
(Fleming 1985 [1958]: 8)

Frame-switches do not discount the possibility that a previously primed frame may be re-primed. If characters remain bound into an unprimed frame, then the potential exists for frame recall. The potential for frame recall diminishes over time. For instance, if a year passes within the text world we might assume that one set of characters are no longer bound into the same contextual frame in which we left them. Similarly, if we happen to be reading a particularly long text, the sheer length of time between one mention of a set of characters and another may reduce our capacity to re-prime the frame into which those characters were initially bound. The author is likely instead to be forced into re-establishing the triggers for the construction of a renewed frame.

Contextual frame theory, then, provides a number of hypotheses about the nature of reading and how readers monitor the progression of a narrative. The concepts described so far, of course, refer to the activities of a reader who experiences no difficulties in processing a narrative. Occasionally, though, a reader may make a mistake, or be intentionally misled by an author/narrator. When this happens, the reader makes a frame repair in order to recover the sense of the narrative in question. Writers often exploit both a reader's capacity for making such misjudgements and their capacity for repairing them in order to generate particular stylistic effects. Emmott (2003a), for example, shows how contextual frame theory can account for the way in which readers react to short stories that incorporate a ‘twist in the tale’, prime examples of which are the short stories of Roald Dahl. Further examples of this kind of technique can be seen in science fiction narratives, ghost stories and fantasy novels. A typical instance of this occurs in John Masefield's classic novel for children, The Box of Delights. In The Box of Delights, the schoolboy Kay Harker is given a magical box by an old Punch and Judy man, Cole Hawlings. The box enables Kay to travel in time, to fly and to reduce himself in size. Its magic is considerable and, as such, it is much sought after by the wicked Abner Brown and his associates, foxy-faced Charles, Joe, and Sylvia Daisy Pouncer. After undergoing a number of adventures in which he manages to keep the box out of the hands of Abner Brown, Kay is finally able to make it to the Christmas Eve midnight service at Tatchester Cathedral. The following example comprises the final paragraphs of the book:

(8)

Organ and brass band struck up, full strength, the Vestry door curtains fell back to each side and out came the great Cathedral crosses and blessed banners with all the Cathedral choir and clergy, voices lifted aloft in ‘O Come all ye Faithful’. By this time the triforium and clerestories, as well as every space in the Cathedral, were packed with faces: all there sang as they had never sung, the singing shook the whole building.
Somehow it seemed to Kay that it was shaking the Cathedral to pieces – Kay himself was being shaken to pieces, his own head was surely coming off, still singing, through the Cathedral roof.
In fact, the Cathedral was not there, nor any of all that glorious company. No, he was in a railway carriage on a bitterly cold day, the train was stopped, he was at Condicote Station with his pocket full of money, just home for the holidays and Caroline Louisa was waking him.
‘Why, Kay,’ she was saying, ‘wake up, wake up. You have been sound asleep. Welcome home for the holidays! Have you had a nice dream?’
‘Yes,’ he said, ‘I have.’
(Masefield 1984 [1935]: 167–8)

At this point in the narrative, the reader will have primed a contextual frame into which Kay and the other members of the Cathedral congregation are bound. Kay is consequently a primed character, as well as the focus of attention as a result of the overt reference to him via the proper noun Kay and subsequent masculine pronouns. The beginning of the third paragraph, however, invokes a frame modification as a result of the narratorial claim that ‘In fact, the Cathedral was not there’. This is contrary information to that which we have so far assumed, and it is difficult to make the necessary modification to the current primed frame. Further confusion is caused by the statement that ‘No, he was in a railway carriage on a bitterly cold day’. This suggests both a temporal and locative frame-switch that is difficult to reconcile, since we have not been provided with an explanation for why the Cathedral was suddenly no longer there in the previous contextual frame. The necessary clarification comes with the revelation that Kay has, in fact, dreamt the whole adventure. As a result of this, the reader is forced to make a frame repair in order to make sense of the earlier statement, ‘In fact, the Cathedral was not there’. This also means that it is necessary for the reader to reappraise all previous contextual frames, and view them as subordinate to the current primed contextual frame in which Kay is woken from his dream.

6.5 Summary and conclusions

In this chapter we have outlined three cognitive theories that deal with various aspects of the reading process. Each of these have pros and cons associated with them, perhaps the most significant of which is the relative lack of empirical testing that each theory has received. Nonetheless, each generates a number of hypotheses which may be subjected to empirical tests, and herein lies their potential value for stylistics. Indeed, some empirical research into these areas is already underway. Emmott, for example, has conducted several experiments into the psychology of reading that have demonstrated support for some of the claims that contextual frame theory makes (see, for example, Emmott 2003b and Emmott et al. 2007). Both text world theory and deictic shift theory are open to empirical testing too, and some of the results with regard to the latter can be found in the volume edited by Duchan et al. (1995).

What should also be apparent from our discussion of these theoretical approaches is that there are considerable interconnections between the three. While text world theory provides a means of accounting for the way in which we construct vivid mental models of a narrative as we read, deictic shift theory accounts for the degree of involvement we might feel in such texts, and contextual frame theory provides an explanation of how we keep track of the complex information that it is necessary to manage in order to achieve these rich representations of a narrative world. It is also possible to map certain elements of these theories onto each other. For example, the notion of a frame-switch from contextual frame theory is roughly equivalent to the notion of a world-switch in text world theory (indeed, Gavins 2007 explains that Emmott's term motivated her choice of the term world-switch as a replacement for Werth's 1999 sub-world). Readers should also be able to note a number of connections between the theories discussed in this chapter and those outlined in Chapter 5. Contextual frame theory, for instance, makes use of several aspects of schema theory in accounting for the frame assumptions that readers have as they monitor contextual frames.

At this point we should also make clear that in our attempts to explicate some of the key ideas in cognitive stylistics, we have made something of an artificial distinction in this chapter and the last between the notions of text comprehension and text processing. Our focus in Chapter 5 was on the means by which readers bring their prior knowledge of the world to bear in the interpretation of texts. In Chapter 6 we have focused primarily on how readers manage the multitude of complex information in a narrative in order to navigate their way through it and make sense of its propositions. However, text comprehension necessarily involves text processing, and vice versa. Contextual frame theory, for instance, involves the management of schematic information as well as more obviously textual information. It is, then, important to realise that the various strands of cognitive activity involved in the reading process cannot in practice be separated out as easily as we might have suggested over the course of this and the previous chapter. Nonetheless, we hope that our presentation of these ideas here has helped to clarify some of the main distinctions between them.

Exercises

Exercise 6.1 Compare the following two openings of novels and consider the differences between them in terms of text world theory. Identify the world-building elements and function-advancing propositions and try to explain the differences in stylistic effect between the two texts. Consider also how deictic shift theory and contextual frame theory might elucidate your analysis.

(9)

Text 1
The summer day was drawing to a close and dusk had fallen on Blandings Castle, shrouding from view the ancient battlements, dulling the silver surface of the lake and causing Lord Emsworth's supreme Berkshire sow Empress of Blandings to leave the open air portion of her sty and withdraw into the covered shed where she did her sleeping. A dedicated believer in the maxim of early to bed and early to rise, she always turned in at about this time. Only by getting its regular eight hours can a pig keep up to the mark and preserve that schoolgirl complexion.
Deprived of her society, which he had been enjoying since shortly after lunch, Clarence, ninth Earl of Emsworth, the seigneur of this favoured realm, pottered dreamily back to the house, pottered dreamily to the great library which was one of its features, and had just pottered dreamily to his favourite chair, when Beach, his butler, entered bearing a laden tray. He gave it the vague stare which had so often incurred the censure – ‘Oh, for goodness sake, Clarence, don't stand there looking like a goldfish’ – of his sisters Constance, Dora, Charlotte, Julia and Hermione.
(Wodehouse 1980 [1969]: 1)

(10)

Text 2
Peter saw her first.
She was sitting on a stone, quite still, with her hands resting on her lap. She was staring vacantly ahead, seeing nothing, and all around, up and down the little street, people were running backward and forward with buckets of water, emptying them through the windows of the burning houses.
Across the street on the cobblestones, there was a dead boy. Someone had moved his body close in to the side so that it would not be in the way.
A little farther down an old man was working on a pile of stones and rubble. One by one he was carrying the stones away and dumping them to the side. Sometimes he would bend down and peer into the ruins, repeating a name over and over again.
All around there was shouting and running and fires and buckets of water and dust. And the girl sat quietly on the stone, staring ahead, not moving. There was blood running down the left side of her face. It ran down from her forehead and dripped from her chin on to the dirty print dress she was wearing.
(Roald Dahl, ‘Katina’ in Dahl 1973 [1945]: 83)

link to answer

Exercise 6.2 Read the following poem and identify the deictic terms and the deictic centres to which they refer. On the basis of having identified these, use the principles of deictic shift theory to describe how a reader is likely to navigate their way through the text world of the poem. What stylistic effects does such an analysis help you to uncover?

(11)

The Rhodora: On Being Asked Whence Is The Flower
In May, when sea-winds pierced our solitudes,
I found the fresh Rhodora in the woods,
Spreading its leafless blooms in a damp nook,
To please the desert and the sluggish brook.
The purple petals, fallen in the pool,
Made the black eater with their beauty gay;
Here might the redbird come his plumes to cool,
And court the flower that cheapens his array.
Rhodora! if the sages ask thee why
This charm is wasted on the earth and sky,
Tell them, dear, that if eyes were made for seeing
Then beauty is its own excuse for being:
Why thou went there, O rival of the rose!
I never thought to ask, I never knew:
But, in my simple ignorance, supposed
The self-same Power that brought me there brought you.
(Ralph W. Emerson, 1834, pub. 1847)