David S. Miall
The history and theory of reading usually takes its beginnings with a standard narrative comprising several well‐known figures: Rosenblatt, the Russian Formalists (Shklovsky, etc.), and the Prague Circle (Mukařovský); and it includes such figures as I. A. Richards, Holland, Fish, Jauss, Iser, and Eco, who (except for Richards) may now be regarded as so many examples of postmodern theory; and a number of later scholars who were responsible for creating and developing an empirical tradition, including Schmidt, Graesser, Oatley, van Peer, Tsur, and Dixon and Bortolussi; the work of these scholars is associated in particular with the organization of the International Association for Empirical Studies of Literature (IGEL; founded in 1987 by Siegfried Schmidt) and their biennial conference. In addition, while the empirical tradition is not yet characterized by a single approach, it is notable for the adoption by some scholars of such methods as stylistics, corpus linguistics, narratology, phenomenology, cognitive psychology, neuropsychology, and evolutionary psychology.
But theories of reading are not an exclusively modern preoccupation, as Robert C. Holub (1984) puts it in Reception Theory: “precursors are not difficult to find” (1984: 13). We should consider earlier thinkers who offered comments on reading without these forming a substantial basis for a theory of response: for example Aristotle in Poetics, Horace’s The Art of Poetry, Longinus’ On the Sublime, or Coleridge’s Biographia Literaria (1983). In Poetics Aristotle is primarily interested in establishing a theory of the formal aspects of literature, drama in particular – his focus is on the play Oedipus Rex by Sophocles, which forms his main example for developing a theory of tragedy. In this endeavor Aristotle makes one observation that has given rise to extensive comment: his mention of catharsis, and its effect on the emotions of the audience (or reader). His account, which “may be considered the earliest illustration of a theory in which audience response plays a major role” (Holub 1984: 13), forms a fundamental part of his definition of tragedy. It is told in a few words: tragedy is “a representation of an action that is serious, complete, and of some magnitude; in language that is pleasurably embellished, the different forms of embellishment occurring in separate parts; presented in the form of action, not narration; by means of pity and fear bringing about the catharsis of such emotions” (Aristotle 2004: 64). It is clear that the reader’s emotions are at issue here. By pleasurable embellishment of the language, Aristotle explains that the text uses verse in one place, song in another. But it is his mention of catharsis that has caused the most debate. What can it mean to bring about catharsis? And are pity and fear obligatory agents in this process, or can other emotions serve?
Many commentators have proposed that the purging of pity and fear is intended, that is, a medical model is being proposed in which, while watching a tragedy, these emotions are eliminated from the system. Other commentators argue that the emotions are not being purged but purified; that is, emotions aroused during the play are being felt in more appropriate ways; for example, an emotion of inadequacy is replaced by an emotion of pride (at a successful performance of dance, for example). Both models have shortcomings. What other emotions may be implicated? To pity and fear, it should be noticed, Aristotle later adds “anger and the like” (2004: 82). What scope does catharsis have? A local effect within a literary text, or a powerful set of emotions determining response to a text as a whole? While we cannot envisage a satisfactory solution here, a shift of perspective to the emotions and their relationships may be helpful. Near the end of Oedipus, for example, we find that one emotion (hubris) is qualified by another (fear), which is in turn replaced by a third (pity). This may suggest the importance of emotions to catharsis: it is often demonstrated in these and other ways for the literary reader, showing that the architecture of a work rests to some degree on the reader’s emotions (as contrasted, for instance, to other models that are primarily cognitive; e.g., text world theory; situation models, etc.). This model, moreover, is not a static one, as Aristotle makes clear in assigning the principal place to the plot: “anyone merely hearing about the incidents will shudder with fear and pity” (Aristotle 2004: 74).
While catharsis provides one perspective on the role attributed to emotions in literary response, Aristotle mentions a second: “language that is pleasurably embellished” through the use of “rhythm and melody” (2004: 64). Such a claim may seem paradoxical given that this passage is describing response to tragedy, yet what he has in mind is perhaps the reader’s admiration for the poetic skill evident in creating an effective tragedy, “the tragic pleasure that is associated with pity and fear” (2004: 74). The tragic poet calls upon the powers of language that Aristotle goes on to describe under the heading of “Diction and Style” (2004: 87). Here one of the key powers of the poet is that of defamiliarization: achieving diction that is both clear and unusual is afforded by the use of expanded, abbreviated, and altered forms of words; the unfamiliarity due to this deviation from normal usages will raise the diction above the commonplace, while the retention of some parts of the normal forms will make for clarity (2004: 87).
Other terms commonly used to signal these effects are “foregrounding”, and “dehabituation.” The first describes a feature of poetic language: certain stylistic devices that stand out for the reader and that may slow down reading and make it more challenging. Other devices include meter, rhythm, alliteration, assonance, and oxymoron. For Aristotle, among many other devices as a poet, “far the most important thing to master is the use of metaphor. This is the one thing that cannot be learnt from anyone else” (2004: 88). The term “dehabituation” is also appropriate since this calls attention to the psychological dimension of the poetic artefact in ways that Aristotle specifies in some detail. A number of theorists mention experiences of dehabituation. For instance, in his Biographia Literaria (1817) Coleridge remembers how his imagination was fired on first reading Wordsworth’s early poems: that it was “the prime merit of genius and its most unequivocal manifestation, so to represent familiar objects as to awaken in the minds of others a kindred feeling concerning them and that freshness of sensation which is the constant accompaniment of mental, no less than bodily, convalescence” (Coleridge 1983: 81). Coleridge here indicates a medical model, where powerful emotions are brought into balance through a “freshness of sensation” so that we come to experience them fully, as though they had never been felt before with such lucidity and clarity.
Longinus’ treatise on the sublime is another important and influential account which has several interesting implications for the reader. The sublime “entrances” its hearers; it “transports us with wonders”; “A well‐timed stroke of sublimity scatters everything before it like a thunderbolt” (Longinus 2004: 114). A sublime passage survives numerous hearings “if it can stand up to repeated examination,” and if it is universal in its appeal, if it “pleases all men at all times” (2004: 120). Among other properties of the sublime is its power to elicit imagery in the reader, that is, when it enables its hearer to imagine that they see something, they “bring it before the eyes of your audience” (2004: 133). Thus in rhetoric, Longinus notes, powerful imagery helps the argument below the surface to succeed: “when it is combined with factual arguments it not only persuades the hearer, but actually masters him” (2004: 135). These comments by Longinus show that the sublime moment exercises a physical presence over the hearer, a disturbance of the body provoking the emotions – all the more effectively for the emotions being hidden, their purpose structured in a state of absorption. Other terms for this experience include being rapt, immersed, captivated, preoccupied: these, among other words, demonstrate the power that may be exercised over the self of the reader by reading. Note that it is commonly the case that emotion is accompanied by imagery, likewise that imagery fosters emotion. This observation suggests how the literary experience may gain pre‐eminence, with emotion capturing the active, expressive self.
My account of Aristotle’s catharsis, Coleridge on dehabituation, Longinus on the sublime, are intended to point to some of the key terms that illuminate the nature of reading, from Aristotle to the present, and to show that these occur in the theoretical discourse on literary reading more generally. A more detailed account would include Spenser, Pope, Wordsworth, Arnold, Ingarden, and a number of other writers with influence in our own time. That influence, however, tends to be identified with the specific name, principles, and theoretical positions of the writer under review. As a result the same names bearing the same roster of achievements tends to constitute the theory of reading, a history that begins early in the twentieth century with the Russian Formalists, then devolves on to the notable figures (in the Western tradition) of Jonathan Culler, Norman Holland, Stanley Fish, and Wolfgang Iser (Freund 1987). In this chapter my intention is to omit another review of the key personae in this history, and to elaborate instead a recent and less‐well known history, drawing on the achievements of the empirical tradition in which the notable figures may (at most) be mentioned.
Aristotle’s account of style contrasts unfamiliar words (deliberately adopted) with normal diction (2004: 87), a distinction that has been regarded as a difference between foregrounded and backgrounded words (van Peer 1986). A parallel pair of terms noted by Mukařovský is automatization and deautomatization. His account continues: “By foregrounding … we mean the use of the devices of the language in such a way that this use itself attracts attention and is perceived as uncommon, as deprived of automatization, as deautomatized, such as a live poetic metaphor” (Mukařovský 1964: 10).
The term foregrounding (aktualisace) is given by Mukařovský’s first English translator (Garvin 1964). As a dimension of the literary experience it has two significant psychological implications. First, in Mukařovský’s account, “the more an act is automatized, the less it is consciously executed; the more it is foregrounded, the more completely conscious does it become” (Mukařovský 1964: 19). In literature foregrounding thus turns attention to the linguistic means, away from the aim of communication; its objective is “to place in the foregrounding the act of expression, the act of speech itself” (1964: 19), that is, the texture of the words themselves and their resonant properties. In this context the aim of the reader is to grasp the implications of the literary style, whether this features phonetic, grammatical, or semantic components, and with their help consider the purpose and implications of the text in question. This is not to overlook whatever communicative purposes the poetic text possesses, as Mukařovský emphasises (1977: 6); but to note that the communicative aim unfolds in the background and may be subsidiary.
The second psychological feature, pointed out by Shklovsky, is to note that the effect of foregrounded features is to “emphasize the emotional effect of an expression” (Shklovsky 1965: 9); a similar point of view is adopted also by Mukařovský: when foregrounding is used, “words and groups of words evoke a greater richness of images and feelings than if they were to occur in a communicative utterance” (Mukařovský 1977: 73). A comment by Coleridge reflects on the power of poetry for the self: “Poetry [is] a rationalized dream dealing to manifold Forms our own Feelings, that never perhaps were attached by us consciously to our own personal Selves” (Coleridge 1962 II: 2086). These comments suggest a link between the foregrounded terms of poetry to the powers of literature for the reader: a point of departure for a theory of literary reading. The comments above allow for the augmentation of literature in feeling, for the enrichment of language in poetry, and for the involvement of the self. But for the moment we turn to examples of empirical studies that have assessed the theoretical claims of critics such as Shklovsky. The account that follows is not, it should be mentioned, a set of instructions for conducting empirical studies, but an informal review of the constructs devised for several of these studies. We begin with a study of foregrounding and related issues.
From the beginning of life the human sensory system of the infant has been sensitive to shifts and changes in the ambient sound – since the first year of life before the acquisition of language – and has related this systematically to the sounds emitted by other humans or itself, for example, when experiencing affection from a caregiver, or the dissatisfaction associated with hunger, or the complex rhythmical patterns that surround us with music. These many sounds are not superseded by the development of language. Indeed, it seems probable that the infant experience when allied to language only develops in complexity: situations that are welded to sounds that repeat each day gain in significance and enable many aspects of the aural environment to be anticipated. Hence the daily round acquires a reassuring predictability, a degree of comfort in which food, other humans and their behaviour, and the sounds they emit, with other regular occurrences, enable the growing child to begin to derive the larger patterns that surround it and that repeat one day to the next.
Having established these patterns it becomes likely that any departures from them will be noticed, and their implications for good or bad will take up a lodging in the infant system either as warnings, or as an array of interesting or pleasant surprises. Some children’s literature is based on the play with language taken to an extreme, as the writing of Dr. Seuss demonstrates here (from his birthday verse):
Today you are you!
That is truer than true!
There is no one alive
who is you‐er than you!
The earliest pre‐reading experience of the child, as we suggested (Miall and Dissanayake 2003), is an important component of the mother–child interaction, commonly known as “babytalk” or “motherese.” It occurs during the first year; and it can be argued that it is the child’s first encounter with a literary or aesthetic experience, an experience that will develop through childhood if the appropriate materials are provided (in print or orally). In time, this will evolve into the challenge of mature literary experience, enjoyed by readers in their teens or 20s and beyond. It provides the basis for the virtually automatic response to the stylistic features of literature – the phonetic, grammatical, and semantic properties already mentioned. Recognition of this phenomenon led us to carry out the first of several empirical studies on literary reading that will now be described.
As noted, a number of literary theorists going back to Aristotle have emphasized the stylistic properties of literary texts, the power of metaphor in particular, seeing our notice of them while reading as spontaneous, recognizing their special powers as foregrounding. Running in parallel are several identifiable psychological phenomena that we might consider as characterizing the literariness of the literary text; that is, a foregrounded feature such as the consonance in the phrase: “damp and cool trickle” (which attracts attention to itself); this correlates with a psychological feature richly presented, such as the sensory properties of a woodland stream. This phrase enriches the narrative context in which it occurs, endowing it with additional properties with the aid of an empirical study. Foregrounded features form an array of properties as mentioned earlier. Our examination of them can be regarded as hypotheses: if a systematic response is made to the foregrounded features by a large number of readers this may be regarded as independent evidence for the autonomous power of the stylistic effects occurring in a narrative or in a lyric poem.
For the purpose of an empirical study, a short story was chosen, short enough to enable it to be read within the time allowed for a laboratory session. As this was a study in stylistics the foregrounded features of the text were identified in advance by the experimenters through a thorough analysis of the text, which contains 1387 words and 84 segments (a segment was defined as one sentence, except where a particularly long sentence was occasionally divided into two or three segments). Our first task was to collate the times taken by readers to read each segment. Our aim was to assess Shklovsky’s claim that in aesthetic response the process of perception is prolonged, that it works to “increase the difficulty and length of perception” (1965: 12). If this was correct, then a segment containing many foregrounded features would take longer to read than a segment of the same length containing few or none. It can readily be seen that this approach is also based on the contrast between background and foreground. Sixty students were recruited for the study, volunteers from senior English classes at the University of Alberta. They read the short story on screen, pressing the space bar regularly to reveal the next segment. Without readers being aware of it, the computer recorded the time taken to read each segment in milliseconds, after adjusting times to segment length and converting to times per syllable.
We were also interested in the judgments that readers make while reading, thus we asked readers to read through the story again by segment, but this time they recorded a judgment of each segment by using a rating scale (from 1 to 5), typing in a number on the keyboard to record their judgment of a given segment. Small groups of readers provided us with ratings on one of several different judgments: strikingness, feeling, discussion value, importance, and imagery (readers’ judgments constituted the dependent variables). For example, readers were asked to report to what extent each “segment arouses feeling in you as a reader, from no feeling to strong feeling.” The ratings of the other judgments were elicited through similar questions. The sets of ratings were then adjusted by calculating the mean for each segment. An index for foregrounding overall per segment was created by computing the mean of the three separate indexes. An index for foregrounding overall per segment was created by computing the mean of the three separate indexes. The data were examined for patterns of relationship, as predicted in particular by the discussions of foregrounding and related issues. A number of relationships were found.
It should be borne in mind that several groups of independent data were created: considering only the data analyzed by segment, there was the foregrounding index and three subsidiary indexes that recorded phonetic, grammatic, and semantic foregrounding. In addition we had the record of reading times per segment, and the judgments of readers on several rating scales that reflected their response by segment. These created a rich and complex model of response to the story segments as a whole. At this point our task was to examine the data for responses that might reflect aspects of such a model; to ask, in a word, whether the stylistic analysis of the story predicted in part readers’ understanding of it.
Are reader’s perceptions lengthened, as Shklovsky proposed, when encountering foregrounded features? Do the reading times of our participants reflect the longer time required to capture the richness and detail of the story segment being read at that moment? If this were shown to be the case, it would be an interesting finding, arguing for the inherent power of the literary text to shape the reader’s experience (a view that was strongly rejected by Stanley Fish, 1980). Several salient findings point to such a conclusion: we find that a simple correlation (Pearson’s) between reading times and phonetic foregrounding was positive, r = 0.377. This result, however, is partly accounted for by factors measured at the same time, such as the other foregrounding components; thus a more appropriate measure is a partial correlation that eliminates these influences; this shows the partial correlation to be, r = 0.301. This second result appears to be less powerful, but there is in fact no important difference between them. In another result the correlation of grammatical foregrounding and segment reading times was shown to be insignificant. The partial correlation for semantic foregrounding and reading times, however, was r = 0.304. The relationship of the rating judgments to foregrounding was analyzed, after conversion to the mean for each rating group: among the judgments examined, strikingness and feeling were significant: partial correlation of the foregrounding index with strikingness was r = 0.297; with feeling it was r = 0.278. Again, this comparison did not represent a significant difference, but the findings in themselves are significant, and showed that for these readers (registering their judgments systematically) a pervasive structure was evoked by the story.
The results shown here go some way to support the claims made for the autonomous power of literary texts – in the case of this work, at least, a modernist short story written earlier in the twentieth century; we later found similar results with several other stories from this period. We can see the text as inviting two ways of reading, a conventionalist and an experiential way. According to Stanley Fish (1980), and several other notable critics, however, we necessarily read within a conventionalist framework. We found that the text contains numerous stylistic features, but that the empirical means employed showed that most readers were unaware of their influence, such as the reading times per segment. Our study has shown that a majority of readers are governed by the same clusters of features in the narrative (for example, the correlation of reading times with the foregrounding index). Fish tells us that a structure of features, designed to determine readings of this kind, operates wherever readers form a part of an interpretative community. This is said to impose on the reader a set of assumptions and conventions in order to make a work of this kind comprehensible. Is this view tenable?
In the previous section (“Miall and Kuiken (1994)”) we considered the capturing of stylistic features overall in a text. It is also possible to examine a text for the presence and effect of a single feature, which may help provide another demonstration of the power of formal features over the experience of literary reading. In the empirical study to be cited next, confirmation of the power and uniqueness of a formal effect is provided, not by statistical means, as shown by the correlations between text feature and the data of reading; now this feature is investigated through the evidence of a pattern of brain waves (ERPs – event‐related potentials) elicited during reading, which occur distinctively in response to the feature in question but not in response to several similar features. The feature in our example is the “functional shift,” the change of an expression from one word class to another. It is demonstrated in this line from Shakespeare’s Othello: “to lip a wanton in a secure couch” (Othello, IV, I, 71), where the noun lip changes function to act as a verb, and “wanton” changes from adjective to noun (Thierry et al. 2008).
Shakespeare’s works contain many examples of function shift, providing the resources for an experimental study of readers’ responses. Different versions of the verbal forms were created by the experimenters, designed to show their distinctiveness through different patterns of brain waves when exposed to participants. Experimental sentences were constructed according to four different categories, each consisting of a set of 40 sentences:
During the testing phase while participants read the sentences in random order, they were asked to judge whether each sentence was acceptable. As they did so, their brain waves were monitored by electrodes and event‐related potentials (ERPs) were recorded. These electrical responses take several typical forms, usually measured in milliseconds (ms), that identify states of the brain in the early stages of a response. Typical responses include the N100 (a negative wave peaking at around one hundred milliseconds in response to a negative stimulus), the N400 (a negative wave caused by a semantic incongruity), or P600 (a positive wave caused by a syntactic violation).
Among the results reported, as shown by ERPs, a response to syntactic violation occurred at 320 ms after word onset—that is, after the critical word was presented. This signals a problem to be processed as early as possible: that is, it alerts us with the opening moments of the N400 response, which typically works to resolve incongruity. Notably the behavioral response, indicating whether a sentence is acceptable, is slower for sentences that are semantically incorrect – a result that anticipates the shifts in reading time that occur at the encounter with foregrounding.
Response to semantic incongruence appeared between 350 and 550 ms; a larger negative wave occurred when a word was semantically unexpected (regardless of its syntactic status). After this faded a second wave of response to syntactic violation occurred at 550–700 ms; reaction times for syntactic violations were longer than for syntactically correct sentences, demonstrating that the functional shift presented a greater challenge to participants’ comprehension. Semantic incongruence elicited a negative wave of 350–500 ms, an outcome with “all the characteristics of the classical N400, a wave traditionally seen as an index of semantic integration difficulty or semantic re‐evaluation.” A high syntactic violation that is also marked by an intact semantic wave, that is, a high P600 but no N400, indicates the Shakespearean function shift – what Davis (2007) calls the P600 surge. The function shift generated the most errors and took the longest among reaction times: it can thus be regarded as a precursor of literary processing. It is an experience that “creates surprise without altering meaning” (2007: 929).
The studies of reading cited so far help to demonstrate the systematic nature of literary response, in contrast to the “whimsical lawlessness of guessing” that E. D. Hirsch (1967: 204) characterized as one major component of the act of reading. Another critic, Jonathan Culler, rejects any socio‐psychological or experimental approach that “would take too seriously the actual and doubtless idiosyncratic performance of individual readers” (Culler 1975: 258). As Bortolussi and Dixon mention, in the light of such comments, many literary critics feared “that the empirical study of real readers would degenerate into sheer interpretive randomness” (Bortolussi and Dixon 2003: 23). These remarks are representative of a prejudice against the empirical among a number of literary theorists. Such observations lead us to wonder why literary scholars would expect such a range of disagreements in the first place when interpreting literary works.
For another perspective I provide two further overviews of empirical studies of reading: an article by Martindale and Dailey (1995) that offers a critical account of Practical Criticism (1929) by I. A. Richards; and Dixon and Bortolussi (2003) on their development of an empirical measure, which they named depth of appreciation.
Current opinion among literary theorists (Martindale and Dailey 1995) is that literary texts are unstable and that readers will thus disagree over their meaning. The article they produce rejects this likelihood. It does so by dismissing the claims of an influential early empirical study by I. A. Richards in Practical Criticism (1929), a study in which Richards had left the field of literary scholarship believing that the student readers he studied “were poor at discriminating between poems, badly in need of the guidance of the experienced literary critic” (Miall 2006: 91). The Richards study can thus be understood as a polemic for the better teaching of literature. That it left empirical studies under a cloud, and laid the basis for the approach of New Criticism, is unfortunate.
Richards’s book is based on the responses of his students at Cambridge to 13 short poems in the form of essays. These responses formed the materials of the study; the poems were presented without identifying the authors. Richards devised a set of ten categories by which he analyzed the responses to the poems: these include perspectives such as the stock response, poetic form, sentimentality, and doctrine in poetry. In this respect, he says, “this book is the record of a piece of field‐work in comparative ideology” (Richards 1929: 6). What he finds most notable is a great diversity of responses to each poem, suggesting that his students demonstrate a serious lack of judgment, an inability to discriminate; they cannot tell the bad poem from the good (1929: 299–301).
In a study of Martindale and Dailey (1995), judgments of the qualities of paintings and music, as measured by rating scales, showed a high level of intersubjective agreement. This led Martindale to ask why literary interpretation would be any different. He conducted a reanalysis and partial replication of Richards’s report. This showed the percentages of students who were said to be favorable, non‐committal, or unfavorable towards each poem. Richards supposed that these percentages were evidence of disagreement. But in a statistical reanalysis of the data of the Cambridge readers a highly significant level of agreement was shown on ten of the 13 poems, and moderately favorable on those remaining. Thus Martindale and Dailey conclude, “Richards seems to have seriously misinterpreted his data at least so far as favorability judgments are concerned” (Martindale and Dailey 1995: 303). As literary critics are well aware, “further empirical studies were probably not done since the case was considered closed” (1995: 303).
He and Dailey then carried out two further studies with the Richards poems. In the first of these studies, 32 participants were recruited from students in introductory psychology; they rated each of the 13 poems on forty 7‐point scales. All the ratings except one showed a high level of agreement. In a second study, in order to replicate Richards more closely, 11 psychology students were asked to read three of the poems and write a short essay (five minutes was allowed for reading each poem, 15 minutes for writing each essay). The essay contents were then scored for Heise’s (1965) norms for evaluation, activity, and potency, which are thought to measure the main dimensions of connotative meaning. Again, agreement was found on all three poems. Thus, they conclude, “people agree in their interpretations of literature to about the same extent regardless of whether they express these interpretations via rating scales or essays” (Heise 1965: 306–7). The problem here for literary scholarship and its history is that the higher level interpretations of interest are far distant from these naive students’ responses; agreement, say Martindale and Dailey, such as Richards’s readers showed, “is not on a level that literary critics care about” (Martindale and Dailey 1995: 307).
The assumption that the interpretive judgments of non‐professional readers are unreliable, even whimsical, has undermined the history of literary reading, with the exception of a few readers or critics. In contrast the following scholars have made notably positive contributions: Jonathan Rose (the historian); David Bleich; Lousie Rosenblatt; Michel de Certeau; Ralph W. Rader; and the Konstanz group led by Hans Robert Jauss and Wolfgang Iser.
Some implications of literary reading were reflected in the studies analyzed earlier for their insights into questions of reading. Among issues for further consideration empirical studies have shown that literary texts are to an important degree stable entities: aside from the pyrotechnics of postmodern criticism, there are several issues arising from such stability; that readers will tend to agree on their interpretations of a given text; that the formal aspects of the text provide a pervasive structure for its interpretation; and that the insights of the ordinary or “common” reader show a degree of continuity with professional readings.
An approach to literariness has been an important component of the studies considered so far, and this is an important part of the last study I describe by Bortolussi and Dixon (2003). For quantifying the literariness of reading the depth of appreciation measure was devised: this helps capture changes in evaluation of a text on a set of measures taken after both first and second readings. In the first study, two texts, “Emma Zunz” (2116 words) by Borges (1983), and a pulp fiction (2752 words) Death Was Her Dowry (Dixon et al. 1993: 18), were read twice by unskilled readers (45 psychology students). On the basis of three ratings questions (out of 15) completed after each reading (Is this good literature? Did you enjoy it? Would you recommend it to a friend?), a marked upward shift in appreciation was found for “Emma Zunz,” but only a minor shift for Death Was Her Dowry. This result shows that untrained readers (as represented by introductory psychology students) were sensitive to literary effects and able to discriminate the differing degrees of literariness in the two texts. Borges was eligible for a literariness upgrade, as it were, but Death Was Her Dowry was not.
In a second study, an interpretation of “Emma Zunz” was proposed based on the theme of unreliable communication in the story. A version of the story was made in which all the passages connoting problems in communication were regularized, being made unproblematic in relation to the communication issue. The same empirical test (ratings judgments) showed that experienced readers (in the unskilled population) showed little upward shift with the regularized version (but a shift in the original version), while inexperienced readers showed no shift with either version of the story. An adaptation of Death Was Her Dowry including passages showing communicative unreliability had little or no effect in producing an upward shift at second reading. This argues that the text feature in itself will have no effect if the overall text is sub‐literary. However, the depth of appreciation measure seems an interesting tool for signifying literariness. Appreciation is, after all, usually a fundamental constituent of the literary experience.
The models of literary reading presented here have involved concepts and methods that are probably quite unfamiliar to most readers. They are models, however, that could begin to refresh a discipline that is widely seen as having outlived its time and exhausted its intellectual resources, and which is currently in some dispute about its aims. It would, among other benefits, require a practical familiarity with scientific theory and method, or at least that part of it appropriate for humanistic inquiries (Gottschall 2008). This would shift attention away from the obscurity of topics that currently preoccupy the time of the professoriate, and redirect it in the service of the majority of the population who find pleasure in reading. As de Beaugrande (1985) observed: “The activities of ordinary readers have not received the attention or respect they merit in view of their social and humanistic importance: the bulk of literature contacts ordinary readers” (1985: 19). This contact comprises an inclusive commitment to several related topics that will illuminate the nature of reading (there may be others). I enumerated these as follows (Miall 2006): studying the ordinary reader; the dehabituating power of literary reading, inviting us to consider the unfamiliar or novel; literariness, arising from the encounter of the reader with an eligible text; the role of feeling, an aspect of the psychology of reading yet to be seriously studied; and experiencing rather than interpreting as the primary focus of literary exploration. These are among the proposals for reconstructing the field of literary studies as I sketched them in a book on literary reading (Miall 2006: 2–3), and introduced them in undergraduate and graduate courses on literary reading. May the models presented in this chapter fire the imagination of present and future scholars.