David S. Miall
Department of English, University of Alberta
A lecture given at the University of Munich, June 25 1998
© 1998 David S. Miall
I make a critical examination of the well-known claim that conventions create the literary object and determine how it is to be interpreted, using the writings of Eagleton and Fish. I claim that literary response is, on the contrary, based in part on some pervasive and important psychological "universals." Some evidence for these is apparent in the systematic effect of formal features of literary texts on the process of reading. With reference to some recent empirical studies in our lab, I will sketch the relevance of this evidence to understanding the response to foregrounding in literary texts and its implications for the self concept of the reader. I will propose that literary studies should focus on the process of response, not on generating more interpretations.
1. The tyranny of interpretation
1.1 Eagleton's dismissal of literariness
1.2 Fish and the predetermined reader
1.3 Against interpretation
2. Laying empirical foundations
2.1 Phonetic iconicity
2.2 A three-phase model of response
2.3 Reading and the self
1. The tyranny of interpretation
It seems to be a commonly held assumption, both inside and outside the academy, that the proper business of the literary critic lies in interpreting and evaluating literary texts. Stanley Fish, for example, proposes that the primary purpose of literary studies is "to determine what works of literature mean"; that our first question should be "What is this poem (or novel or drama) saying?" (Fish, 1995, p. 25). But I will argue that the dominant focus on the meaning and value of literary texts, interesting though the results may often be, has served to misrepresent the nature of literature. I will suggest that the principal question behind such efforts -- What is this text about? -- is a misleading place to begin. Literary texts are not read primarily in order to extract, record, or elaborate the ideas they contain. The fact that the critical literature on any given canonical text is usually replete with disagreements, oppositions, and contradictions is one sign that interpretive ideas do not warrant the status attributed to them and that the interpretive enterprise may actually be engaged in something other than the aim of understanding and assessing literary texts.
Those who have looked in on literary studies from the outside, such as Jean-Claude Gardin (1991) or Herbert Simon (1994), have been troubled by the ceaseless proliferation of interpretive studies. Who needs yet another account of the meaning of Hamlet or Wuthering Heights? Yet, out of the current sense of crisis in literary studies, the prescriptions for change that appear most influential are primarily concerned with proposing more interpretation, not less, a phenomenon that I will examine more closely in the writings of two influential theorists, Stanley Fish and Terry Eagleton.
The most troubling implication to be found in contemporary theory, however, is its refusal to consider literary reading as a real event involving real readers. While the theorists who now largely dominate the field of literary studies are, of course, readers, their interests are by no means representative of what non-professional readers experience in their encounters with literary texts. I will argue that the study of ordinary reading, through a reliance on empirical methods, now seems essential, if literary studies is to maintain its validity as a discipline (it is already being replaced in many quarters by one or more of cultural studies, discourse analysis, media studies, etc.). But real readers, as you probably know, are of no theoretical interest to the present generation of theorists. A central reason for this neglect can be found in a representative set of arguments by Terry Eagleton.
1.1 Eagleton's dismissal of literariness
In the course of his influential book, Literary theory: An introduction (1983), Eagleton attempts to replace the concept of "literature" with that of rhetoric; literature becomes merely another form of textual practice. This appeal to rhetoric is perhaps the commonest theoretical maneuver currently being made by literary theorists. It is an approach that, despite its claims to political relevance, turns literary studies into another form of interpretation: given these textual structures, it suggests, and this social context, these are the meanings we will find in a given text. This approach arrogates to itself the power to decide what a text means and how it will be read, as Eagleton's account makes clear. The appropriate method of study will be, in his words, a "concern for the kinds of effects which discourses produce, and how they produce them." Rhetoric, from ancient Greece onwards, "examined the way discourses are constructed in order to achieve certain effects" (p. 205). Rhetoric, he notes, "shares with Formalism, structuralism and semiotics an interest in the formal devices of language, but like reception theory is also concerned with how these devices are actually effective at the point of 'consumption'" (p 206).
This last phrase might seem to connote the study of reception, an interest in the response processes of actual readers, but it is clear that this is not what Eagleton has in mind. For him, to study rhetoric is to interrogate textual strategies in the light of certain ideological presuppositions, in brief, to see "discourse as a form of power and desire" (p. 206). Having already dismissed reception theory as a theoretical enterprise earlier in the book, in his account of Fish and Iser (whose readers also exist only in theory), Eagleton is not about to turn to the empirical interface of texts and readers. It is clear that for him it is inconceivable that we might find out if texts are, in his words, "actually effective" by asking readers.
Although Eagleton reserves his strongest criticism for the liberal humanist, whose view of individual freedom is said to be an abstraction (p. 207), Eagleton himself grants no real authority to readers. The readers who figure in his book remain imaginary ones, created in the light of Eagleton's political premises. As an answer to the undemocratic practices of prior formalist modes of literary study, which Eagleton rightly challenges, this is really no answer at all. It is to exchange one form of authority over interpretive practices for another. Literature, as Eagleton insists, "has a use" (p. 208); but we will know nothing of real value about that use until we learn how to ask real readers what uses literary texts have in their lives.
But Eagleton's bias is typical of current approaches to literature: it exemplifies what we might call the Received View of literature that now dominates literary studies. In this view it is cultural norms that determine what will be found literary; such norms are the product of the prevailing ideology, they depend on class, race, or gender interests, and they bring the literary system into being. Thus there are no properties that are distinctive to literary texts -- neither text properties that endow certain texts with "literariness," nor any abiding or transcendental truths visible through what literary texts represent. A common term for this stance is anti-foundationalism. In the second example of it I will now discuss, we find Stanley Fish arguing against what he calls "foundationalist" theory, which he attributes to Romantic writers (such as Coleridge and Shelley) or the New Critics.
1.2 Fish and the predetermined reader
In brief, Fish's argument is that foundationalist theory "will never succeed" because "it cannot help but borrow its terms and its contents from that which it claims to transcend, the mutable world of practice, belief, assumptions, point of view, and so forth" (Fish, 1989, p. 321). Fish seems to believe that the human world is necessarily excluded from any determining foundations, because all human activities are social practices, wholly constructed out of socially determined beliefs. For Fish there is no realm below, or apart from belief. A belief, he asserts, "is a prerequisite for being conscious at all" (p. 326). Thus, while we undoubtedly experience such biological "facts" as birth, death, hunger, sex, and the like, we have access only to our beliefs about them; there can be no direct experience of them. There is, he tells us, "no way of testing our beliefs against something whose source is not also a belief" (p. 322).
It is within this framework that Fish argues the impossibility of intrinsic or formal properties in literature. Just as meaning cannot be computed from the properties of language, Fish says, neither can literary meaning be the product of formal features of texts. Formalism is construed narrowly by Fish as a way of guaranteeing textual meaning. As I will now suggest, the key limitation of Fish's view is his inability to conceive of other kinds of formalism, or to separate formal properties as agents from how they might be construed by readers. Fish's main argument are contained in his essay "Going Down the Anti-Formalist Road" (Fish, 1989).
Fish argues that there are no formal properties of language that serve to create meaning. No formal relations, including grammaticality itself, are predictable. Thus formalism has to be rejected in total: "once you start down the anti-formalist road," Fish says, "there is no place to stop; remove the connection between observable features and the specification of meaning, and you also remove everything else that is supposedly independent of context" (p. 2). Since we do use language with generally predictable results, however, it follows that speaker context determines meaning, that is, the interpretive community in force at any given time. Everything stable about language meaning comes from outside the formal properties of language itself. The problem with this view is the everything: i.e., Fish's totalizing of the argument.
Fish's strategy here is typical of his way of arguing. Since the distinction between formal and contextual aspects of language is untenable, any formal aspect must be illusory, attributable to agreement of speakers and hearers in a given context. Fish's strategy is to collapse oppositions, as John M. Ellis has noted. What Fish does, says Ellis, is
find a pair of opposed concepts; show that one pole is not completely distinct from the others; pronounce the opposition an illusion; then conclude that they are really both the same and that there are no important differences between things that originally seemed distinct. (Ellis, 1997, p. 174)
Or as Fish himself has put it, "One deconstructs an opposition not by reversing the hierarchy of its poles but by denying to either pole the independence that makes the opposition possible in the first place" (p. 211).
Of course, the pure formalist and Fish (the contextualist) are both mistaken. Language contains both formal and contextual components, as almost any example of spoken or written language will show. The fact that the referents of nouns or verbs are speaker dependent, for example, doesn't deprive nouns and verbs of their specific functions in supporting the grammar of a sentence, a claim that Fish's argument explicitly denies. All well-formed sentences contain a topic noun and a verb: this is a statement about the formal properties of language that underlie meaning. Similarly, sentence construction invariably follows the topic-comment, or given-new structure; again, this is a formal component that underwrites the relation between contributing words. Sentences are often structured deictically, that is, speaker or topic situation are determined by formal markers. Thus, we can show the occurrence of a number of formal components, some of which appear to be universal, occurring in every language studied; other formal properties are specific to languages of a particular group. In any case, it is obligatory for a competent speaker, regardless of context, to employ such formal components in computing the meaning of a given utterance and for a hearer to use them in understanding the utterance.
Fish's anti-formalist road is, for this reason, a misleading figure, since it implies that once we start down the road away from the formalist enclave (as Fish says we must), there is no return; either we accept the formalist proposal (and live in illusion), or we see that formal features cannot exist. But a more appropriate figure would be a field, or a two-dimensional graph, with formal features forming one dimension and contextual features forming the other: depending on which part of the field you choose, formal features may play a larger or a smaller role, but there is no part of the field in which formal and contextual features do not co-exist, jointly determining the interpretation of sentence meaning. As I will show in a moment, the understanding of literary texts appears to be determined in the same way: literary meaning is the joint outcome of formal properties and the contingent, context-bound construals of individual readers.
For Fish, however, theory itself is only another name for a set of practices derived from a given set of interests. (I will not mention here the grotesque view that results when Fish turns to discuss the law.) For Fish, "theory" only has the power that local practices give it: "when theory has consequences, they will be rhetorical, not theoretical" (p. 14-15); theory, he adds, is "one among many rhetorical forms whose impact and sway are a function of contingencies . . . it can neither predict nor control" (p. 25). But Fish stacks the evidence: since he never offers examples of theory other than than those deriving from socially contestable situations, such as carefully chosen law cases, the term "theory" does no work not accomplished by some already existing set of social practices. This restriction makes his general conclusions about the nature of theory and its implications for the claims of formalism less than conclusive; in particular, it enables us to see how Fish comes to overlook and disregard the formal properties of literature that make a difference to readers.
This problem is shown by the difficulty Fish has in explaining how a social constructionist approach allows beliefs to be formed or changed. According to Fish, all we do when we change our minds is to change one belief state for another. He disallows partial or incomplete states of belief, as though a belief was necessarily an all-or-nothing matter. Thus, the possibilities of doubt, hypothesis, being tentative, uncertain, questioning, undecided, sceptical, or ambivalent, all appear to be impossible in Fish's world. He appears to hold this position, because the opposite of belief is defined not as unbelief or doubt, but as "aperspectivity," standing in the light of the truth. Since this is impossible in Fish's view, only belief states are possible, in which one (pervasive) belief state is exchanged for another. He characterizes belief as a power, never relinquished, "to see the world from a point of view in relation to which other points of view are objectively (obviously) false" (p. 19). Degrees or shades of belief are impossible in Fish's monochrome world: there is no way, he insists, in which you can loosen the hold a belief has upon you (p. 21). He finds it impossible that "one can believe an interpretation and not be convinced of it . . . [or] a two-stage process in which one first has a belief and then must determine whether or not to believe it" (p. 115). This view leads Fish to the claim that literary texts cannot be indeterminate, a position which makes it impossible to understand some central features of the literary response process.
Fish decides that indeterminacy in literary reading is impossible, but does so by once again promoting an illegitimate contrast. Perception, he tells us, "(and reading is an instance of perception) always occurs within a set of assumptions that preconstrains what could possibly be perceived (or heard, or tasted, or touched)." If this were not the case, "readers would have to be in a position to specify significance in a random or irresponsible way." That is, if reading were not determined at every moment, as Fish insists, its indeterminacy would be arbitrary and whimsical. Thus it also follows that the subjectivity of individual readers is an illusion, since "the observer is never individual in the sense of unique or private, but is always the product of the categories of understanding that are his by virtue of his membership in a community of interpretation" (p. 83). In this way, Fish once again disallows the possibility of doubt or ambivalence in understanding, a position that he caricatures as "a kind of dead space when one has only the words and then faces the task of construing them" (Fish, 1980, p. 318). On the contrary, says Fish, interpretation is itself "a structure of constraints, a structure which, because it is always and already in place, renders unavailable the independent or uninterpreted text and renders unimaginable the independent and freely interpreting reader" (p. 98)
What is wrong with this picture is Fish's insistence on totalizing it, once again, to include every aspect of the act of interpretation. While he is right to suggest that reading is never arbitrary or whimsical, and that there can be no such entity as a "freely interpreting reader," his insistence on the total predetermination of the act of reading by socially-given constraints would make the act of reading (literary or otherwise) a mechanical, automated procedure, a process of "reading off" rather than reading. In brief, Fish's arguments fail because they create and then collapse oppositions that are essential for understanding literary response as a process. The opposite of the socially determined reader is not an arbitrary, irresponsible reader; the opposite of the socially determined interpretation is not a fixed, essentialist meaning; the opposite of a state of totally determined belief is not a state of non-belief. Fish's account makes inconceivable why anyone would become a reader of literature, and it makes incomprehensible the existence of the literary, as a form distinct from other types of text; and it renders invisible the challenge that literary reading presents to the reader's prior assumptions and self concept. On each of these grounds, an alternative perspective based on empirical study is required, and that is what I will outline in the second part of this talk.
1.3 Against interpretation
But the question "What is this poem (or novel or drama) saying?" is not the place to start. Generating more interpretations is not the way to understand reading. Literary texts have evolved to address their readers through action not ideas; their primary vehicle for this purpose is feeling. An interpretation can never be more than a distant, possibly distorted echo of the reading experience. The action at stake in reading is not usually direct but potential, encoded in feeling: feeling situates readers in relation to complex modes of experience, memory, and social understanding; as such, literature intervenes on readers' modes of feeling and may challenge them or modify them in significant ways. Such challenges lead primarily to action, not thought: they embody alternative ways of construing the self and its position and effect in the world. But feeling also has another important property: it offers a distinctive set of processes that shape response to the formal aspects of literature. It does this through its self-referential aspects, and by its power to provide alternative perspectives that cut across existing cognitive domains. Feeling, in other words, combines both universal and culturally-specific features, and we need to take both into account in attempting to understand the response to literature. It is this understanding that provides the main basis of our empirical approach to literary reading. As you can see, this results in a quite different emphasis on reading than current, postmodern accounts.
But my intention here is not to propose another literary theory to act as an interpretive engine, particularly not by positioning myself in relation to current theory or offering an extended critique of its problems. This merely weighs one rhetoric against another, an enterprise that seems to me inherently inconclusive. Like the combatants on the western front in the First World War, a little territory is gained one way, only to be lost again, with much expenditure of lives and little actual progress to show. Moreover, literary theorists as a rule have been content to remain at what I would define as a pre-theoretical level: literary theories cannot be right because they cannot ever be wrong. There is no evidence that could confute a literary theory, thus such writings are strictly speaking no more than interpretations. Literary theorists, like Galileo's inquisitors, typically refuse to examine evidence in the empirical sense; offered a telescope, they rule that such an instrument cannot exist, or that it exists only as an ideological construct rather than as a tool to aid perception.
I have a rather different kind of literary theory in prospect, one with two unusual and distinctive features. First, it would be consistent as far as possible with what is known in the social sciences and natural sciences; second, the theory will be framed in ways that make it amenable to empirical testing. Empirical research on reading, then, must be seen as the centre of a new approach, one that holds out the possibility of creating a durable paradigm for future literary studies. If we cannot understand the role and functions of literary reading as it occurs in the community at large, after all, then we risk bringing all our understanding to nothing.
2. Laying empirical foundations
The Received View of literature proposes that "literariness" is an illusion: there are no formal properties that characterize literary texts. All we have are interpretations of texts, and these only exist within a contest of beliefs, a realm of unrelieved relativism. In answer, I have suggested that literary meaning is a combination of both formal features and contextual meanings imported by readers. I have also suggested, in opposition to Fish, that meaning may be indeterminate: the meaning of a text, in other words, is emergent; it requires some interpretive work by the reader. And I have also suggested that literary reading challenges and modifies the reader's self concept, and that the main vehicle for this process is feeling.
I would like now to sketch three examples of empirical research that tend to support these claims. I will first outline my current work on phonetic iconicity and show how this influences readers' responses to literary texts: it provides an example of the inherent power of formal features in texts. Second, I look at the process we have identified through which readers experience and attempt to resolve indeterminacy during reading, what we have called the "three-phase" model of response. Finally, I will look at one specific example of a reader's comments during reading to illustrate the challenge that a literary text can present to the reader's self concept; and I will suggest that some aspects of this example are typical of what we might expect of literary reading in general.
2.1 Phonetic iconicity
Do the sounds of words convey meaning? This is the iconicity hypothesis. The most familiar form of sound-meaning is onomatopoeia, the imitation of a natural sound by a word form: for example, the words hiss, miaow, or crack each appear to embody phoneme clusters that sound similar to the event they name. In Hugh Bredin's recent study of this phenomenon, he suggests that "onomatopoeia is not a trivial and incidental phenomenon of usage, but answers to a deep-seated need that lies at the heart of the linguistic consciousness. We want language to be onomatopoeic" (Bredin, 1996, p. 560).
But onomatopoeia is only one small and unsystematic dimension of phonetic iconicity. It was the dimension that Saussure was able easily to dismiss, arriving at his influential claim that the sound of language is arbitrary, that is, conventional (1974, 67-70). Other, potentially more significant forms of iconicity have been investigated by various linguists in several different kinds of empirical study, as Roger Brown showed some years ago in his review in Words and Things (1958, Chapter IV). Some studies have examined contrasts between phonemes, where participants are invited to classify as light or heavy pairs of items, such as the nonsense words mil and mal. A surprising degree of agreement is obtained (e.g., Sapir, 1929). Pinker and Birdsong (1979) examined several phonetic features in a study of word order preference: these include fewer vs. more syllables (Panini's Law), short vs. long vowels, less vs. more obstruant consonants, etc. Their empirical findings showed strongest support for Panini's Law and vowel quality (high vs. low).
In none of these studies, however, has phonetic iconicity in literary texts been examined using empirical methods. This is what I have now begun to do, by developing several measures to represent particular formal properties of phonemes. These are of two kinds. There are the relative properties, such as light vs. heavy, or soft vs. hard, given by virtue of the position in the oral tract at which a phoneme is pronounced. Second, there are the fixed lengths of the vowels: either absolute length, which involves identifying such long vowels as [a] in bard, and [uw] in food, and all the diphthongs; or the vowel lengthening that occurs before voiced stops and fricatives, known as the vowel shift.
Thus I created measures based on several different experimental orderings of phonemes. For vowels, pronunciation in English depends on two dimensions of tongue position: from front to back and from high to low (see, for example, O'Grady, et al. (1989), pp. 28-31). The twenty phonemes of English were weighted in these two orders (the front to back ordering was independent of the high-low position; the high-low ordering included the front to back order at each level). The 24 consonants were ordered from soft to hard, with a front-back ordering at each level, and voiced consonants placed before unvoiced. I also produced a combined vowel and consonant measure by weighting all phonemes according to their front-back position; this measure was termed "presence," indicating the supposed egocentricity of phonemes pronounced in a forward position. Lastly, I created a measure for the frequencies of each type of fixed vowel length (that is, absolute vowel length, and vowel shift). I employed SAY, the internet version of rsynth, to obtain a phonetic transcript of the words and texts I examined, then coded the average weight of the phonemes in each word or text segment using the various measures I created. This enables me to show, for example, that a given line of poetry contains a preponderance of light vowels and soft consonants, or that it contains an unusually large proportion of long vowels.
I used these measures to make a study of the phoneticity of two portions of Paradise Lost. I compared samples of text from Book II dealing with Hell (Satan's encounter with Sin and Death at the exit from Hell: II.629-814) and from Book IV offering the first view of Eden (IV.205-355). If phonetic iconicity is drawn upon to create an underlying tone or mood, then one might expect these two passages to contrast systematically in their use of vowels and consonants: the confinement of Hell should be reflected by a greater proportion of narrow vowels and hard consonants than Eden, which offers a sense of light and space. Using the mean rank ordering of phonemes to measure differences between the two passages (producing a mean weight for each line of text), the Hell passage was found to contain significantly more front than back vowels (i.e., the narrow sounds connoting confinement), t(335) = 2.80, p < .01; at the same time it contained significantly more hard consonants, t(335) = 2.48, p < .02. In Book II the line with the highest number of hard consonants was 714: "Each cast at th' other as when two black clouds." Lines with a high frequency of soft consonants in Book IV are 207, "In narrow room, Nature's whole wealth, yea more" and 260, "Luxuriant; mean while murmuring waters fall"; and in Book IV a line with a high number of wider, back vowels is 256, "Flowers of all hue, and without thorn the rose." On vowel length, the vowel shift measure showed significantly more longer vowels in Eden in comparison with Hell, t(335) = 1.822, p < .05 (absolute vowel length was similar in both passages).
If phonetic differences are as systematic as this example suggests, then we might expect to find effects on reading. To study this, I examined phonetic influences on readers of five literary stories for which I already have data on foregrounding and several reading measures. The five stories had already been coded for phonetic foregrounding, using trained judges to count phonetic features that they thought likely to be noticed by readers. Second, mean reading times per segment were provided for all five stories (and in each case had been found to correlate with foregrounding). For three of the stories, mean ratings for affect and other features had also been collected. The phoneme measures were then correlated with each of the story or response variables, as shown in the following Table:
Table. Phonetic influences on response to literary narratives:
Correlations of phoneme measures with story and reading dataO'Faolain Woolf Mansfield Ramos Wilson (df 82) (df 75) (df 84) (df 283) (df 273) ---------------------------------------------------------------------------------------- vowels vowel length phonetic fg -.524** .128 .082 .065 -.145* reading times .533** .341** .003 -.018 .319** affect -.109 -.051 .007 -- -- vowel shift phonetic fg .328** .080 -.044 .012 .137* reading times .308** -.057 .009 .116 -.129* affect -.133 -.079 .165 -- -- front-back phonetic fg -.263* .188 -.268* .090 -.120 reading times .335** -.064 -.423** -.086 .175** affect -.191 -.219* -.098 -- -- high-low/front-back phonetic fg -.271* .104 .258* .165* .069 reading times .312** -.196 .017 -.116 -.071 affect -.103 .263* -.121 -- -- ----------------------------------------------------------------------------------------- consonants soft-hard phonetic fg .253* -.043 -.342** .377** -.010 reading times -.297** -.310** -.328** -.419** .004 affect .291** .109 .051 -- -- front-back phonetic fg .340** -.082 -.080 .368** -.112 reading times -.018 -.056 -.206 -.334** .099 affect .192 -.123 .026 -- -- ----------------------------------------------------------------------------------------- presence phonetic fg .121 .038 -.230* .370** -.157* reading times .204 -.091 -.425** -.338** .179** affect .040 -.247* -.039 -- -- ----------------------------------------------------------------------------------------- *p < .05 **p < .01Table notes. 1. Vowel weight was analysed in two forms: a front-back distribution (with tongue forward given the highest score and tongue back the lowest), and a high-low distribution (oral cavity narrow scored the highest to wide scored the lowest) with front-back at each level. 2. Consonants were weighted in two forms: from soft (highest score) to hardest (lowest score) with a forward-back distribution at each level, and voiced before unvoiced versions of a phoneme; and from front to back. 3. Presence signifies all the phonemes weighted by a forward to back position. NB. The data analysed, except for affect ratings, was converted to an index by syllable for each story segment.
These results indicate that phonetic tone influences the reading of all five stories in one way or another. The differences between the stories also suggests that each has its own phonetic style which impacts reading differently. For the O'Faolain story, "The Trout," longer reading times coincide with vowels that are both more frontal and higher and with the harder consonants. Reading times for the Mansfield story, on the other hand, appear to be more influenced by back vowels in combination with harder consonants. In fact, in four out of the five stories studied, the harder consonants appear to have contributed to longer reading times (perhaps the fricatives and plosives characteristically take longer to pronounce), whereas vowels have a more variable influence.
The phonetic foregrounding index also shows a variable relationship with phonetic weighting, coinciding with back or low vowels and softer consonants in the O'Faolain story, but back and high vowels and hard consonants in the Mansfield. This shows that foregrounding, as far as we have defined and analysed it (Miall & Kuiken, 1994), has a role independent of overall phonetic weight. Phonetic weight can be seen as contributing to the background style of a given writer, whereas foregrounding occurs when specific local tonal effects are sought, a view that tends to argue in favour of Mukarovský's claim that stylistic effects in literary texts are systematic and hierarchical. These results suggest that O'Faolain's story achieves its phonetic foregrounding effects against the overall phonetic coloration, whereas Mansfield's are gained rather by intensifying it. Readers' affect ratings also point to differences between the feeling tone of stories that readers are aware of, since of the two significant correlations obtained, readers of the O'Faolain story appear to have been influenced by the softer consonants, whereas Woolf readers attributed more feeling to segments containing back or high vowels. (The absence of significant correlations with the phonetic foregrounding index for the Woolf story is not unexpected, since in this story phonetic effects were found to have no appreciable effect on reading times; response to foregrounding was primarily due to grammatical and semantic deviations.) The presence measure seems to show readers of the Mansfield and Ramos stories lingering more over segments connoting distance rather than presence, whereas reading times for the Wilson story appear more positively influenced by presence.
2.2 A three-phase model of response
Phonetic iconicity is one example of a formal feature that, as I have suggested, appears to influence reading systematically. It would appear to contribute to readers' tonal awareness, and thus to form one strand of the local background peculiar to a given text. It is against this background, in part, that specific local stylistic features will be recognized. The next issue to examine, then, is what occurs when readers encounter phonetic or other features that they find striking? How do they take account of them?
Following the accounts of the Romantic writers and the Russian Formalists, we have proposed that the response to striking stylistic features in texts (or foregrounding, in Mukarovský's term), is defamiliarization. When readers encounter a phrase with some notable alliteration or a metaphor, for example, we suggest that this may cut across the text meaning currently being developed by the reader: in other words, to put it schematically, the familiar world view in which the reader is operating is brought into question at such moments. We have found evidence of several components of the response of readers during the encounter with foregrounding: their reading tends to slow down, they report a sense of strikingness, and they are aware of a greater degree of uncertainty about text meaning. But perhaps the most notable feature accompanying such moments is the awareness of heightened feeling. We reported these phenomena in a paper four years ago (Miall & Kuiken, 1994). In that paper, however, we did not go on to examine what the consequences are of these components -- although we offered some speculations that have now begun to find support in some further studies. The principal question concerns the role of feeling.
I had already noticed some years before (Miall, 1988) that the time taken to read the segments of a literary text and the degree of feeling that readers reported show a systematic relationship, but this relation is not a linear one. Looking at responses to two stories, I found that at the beginning of a story episode, greater feeling tended to accompany longer reading times. During the later part of an episode the reverse occurred: greater feeling accompanied shorter reading times. How is this to be explained? I suggested that when a new episode begins, readers are searching for an appropriate context in which to understand it. Feeling provides the main vehicle for this search, thus when something in the text prompts heightened feeling, readers dwell a little longer on such passages in order to interrogate their feelings for clues to meaning. Once a context has been established, however, feeling provides a way of developing a meaning already in place; thus heightened feeling now makes reading more efficient, as shown by its co-occurrence with shorter reading times. In this way, later in an episode, we see readers recontextualizing what they had earlier found defamiliarizing.
In our more recent work, I have attempted to refine this account by looking more carefully at the determinants of this non-linear relationship between reading times and feeling. Having found that heightened feeling occurs mainly in response to foregrounded passages, where we also see longer reading times, it becomes possible to use these moments as starting points and to examine what occurs downstream from them during the reading process. The method I devised is unusual, but has revealed some interesting and systematic findings, and it has led us to propose a three-phase model of response to foregrounded moments. The process appears to work like this:
Phase 1: Defamiliarization. The reader encounters a striking passage and experiences defamiliarization, which arouses feeling. Evidence for this comes from higher ratings for strikingness and uncertainty; and the association of increased feeling with longer reading times.
Phase 2: Search for context. Through the feelings aroused by foregrounding the reader works towards retrieving experiences, memories, or ideas that will provide a new context for understanding the story so far. There is a higher positive correlation of affect ratings and reading times, signifying the role of feeling in searching for a new context
Phase 3: Recontextualization. Having found a context for understanding the story, the feelings involved in this process now make reading more efficient. At this stage we see increased feeling associated with shorter reading times. In a parallel remindings study, readers report a greater frequency of shifts in story understanding (11-12 segments following the initial encounter with foregrounding).
In the responses to the short stories we have studied, a series of patterns corresponding to Phase 1 to Phase 3 responses can be seen. When successive correlations of reading times and affect ratings over six segments (time/affect correlations) are computed, an alternating cycle of positive and negative correlations is obtained. Here is the pattern for one of the stories we have studied:
Figure 1: Correlations of reading times and affect
This pattern can then be compared with the successive mean foregrounding (the means across six segments). The two data sets correlate, but do so most strongly when the time/affect data is shifted three segments back in time: r(74) = .366, p < 001. In other words, the effect of foregrounding on the phasic pattern of response shows an average time-lag of three segments, pointing to the relatively long retrieval time required for feeling to initiate the new interpretive context.
In the following graph, data has been averaged across the three main cycles that appear to occur in "The Trout"; I have also added the mean data on the frequency of story shifts from the Remindings study:
Figure 2: Three Phases of Response to Foregrounding
This pattern reminds me of Coleridge's (1817/1983, ii.11) comment on literary reading. He described the reader's journey through a literary text as like a snake: "at every step he pauses and half recedes, and from the retrogressive movement collects the force which again carries him onward." If we take the time/affect graph as the marker of this process, the curve above the line signifies the pause (readers are reading more slowly while they experience heightened feeling); but this gives way to the curve below the line, when reading is being carried forward more rapidly in the light of what feeling has put in place.
So far, then, in the research I have described, I have pointed to several important formal influences on the reading process: the specific phonetic tone of a literary text, with its measurable influence on the reading times and affect ratings of readers; the response to foregrounding, with its characteristic consequences dependent on feeling. You will also notice, however, that I have said nothing about interpretation. Clearly, readers are interpreting texts, but what I have shown you so far are the formal processes through which interpretations appear to be constructed. It may be the case that what is most distinctive about literary texts are these formal processes that they initiate, not the interpretations to which they give rise. That is, the question which Fish finds so important, "What is this poem (or novel or drama) saying?" (Fish, 1995, p. 25), may be of less interest to readers than the process they experience during the act of reading. If defamiliarization is a key component of literature, perhaps the primary function of reading is to challenge our cognitive and emotional stereotypes and enable us to rethink them. Literature, that is, retunes or modifies the ways we think; it raises questions, rather than providing us with new wisdom.
2.3 Reading and the self
What I will show you now is a specific example of a reader's comments during reading. This will provide a specific example of the kinds of thinking that may typically occur during reading, when readers are reminded of issues concerning their particular concepts of the self. This response is the last of five comments elicited during a reading of Coleridge's long narrative poem, The Rime of the Ancient Mariner. Readers were invited to note passages that they found striking, then after reading the whole poem they returned to selected passages and made detailed commentaries. My remarks will be brief and superficial, but will show how the thinking here coincides with the sequence of formal features I have already described.
The lines of the poem selected by the reader occur quite late in the poem:
Like one, that on a lonesome road
Doth walk in fear and dread,
And having once turned round walks on,
And turns no more his head;
Because he knows, a frightful fiend
Doth close behind him tread. (Mariner, 446-451)
Identifiers: scared, danger, lonely
I seem to be picking on a bit of a theme of threatening.
Description: I'm just going to share the emotion of being alone, in the dark, with this threat. Knowing that there's nothing you can do about it, keeping on walking and pretending it's not happening, just because there's no other way to cope with it, you can't run from it. All you can do is hope that somehow or other it magically just disappears and leaves you alone. Loneliness is being, having no one around to help you, feeling like you're the only person because, there's nobody else that's near enough to do you any good, so when this being decides to grab you, it doesn't matter if there's 40 million other people in the world, if none of them are there to save you from this. I also sense there's no point in fighting this because, like it's a guilt thing, he's the one that's responsible for what's happened, he's the reason that this thing is following him, so there is no point in trying to get away from it because, it's your fate. It's just a bit of a reminder that everybody dies. Whatever's following him is going to get him. You don't know how long it's going to go and you don't know when it's going to get him, but you know that eventually that it will.
The heightened feeling of the reader at this point seems clear from her opening comment, "I'm just going to share the emotion of being alone." But the feeling also resonates with some underlying concept about the self, shown by the use of the pronoun you: " Knowing that there's nothing you can do about it." The words you or your occur 13 times during this comment. While it is a pronoun that can be used defensively, to distance the self from some state or feeling, here it appears to be used as a way of enlarging a specific personal feeling of which the reader has been reminded ("I'm just going to share"). She finds herself led to inquire into the meaning of this feeling with the help of the poem: "Loneliness is being, having no one around to help you."
After exploring this feeling, she returns to the situation of the protagonist: "it's a guilt thing, he's the one that's responsible." Thus, we can see unfolding a set of processes similar to the first two phases I set out earlier: a first awareness of a feeling, with some personal relevance; the use of this feeling to locate a meaning for the poem, which is then applied to thinking about the position of the protagonist. Finally, in what is perhaps the most interesting part of the commentary, we see a convergence of the protagonist's situation with that of the reader: the "he" and "you" appear to become interchangeable. Although "this thing is following him," "it's your fate." The story understanding that emerges at this point appears to be "everybody dies." While this is certainly not a profound insight in itself, the way in which it is reached has made it personal to the reader, and enabled her to pursue a particular theme that seems to have concerned her throughout the poem, as she realizes in her first comment "I seem to be picking on a bit of a theme of threatening."
Thus, across the course of the commentary we can see elements of the three-phase model that are initiated by feeling and that lead in time to new understanding of the text. What is particularly distinctive here are the signals that feeling has implicated the reader's self-concept and led her to identify the protagonist's predicament in some way with her own. To be included in the fate "everybody dies" is no longer (at that moment) a stereotype, what everbody knows. It has the quality of something felt, experienced, a knowledge that is being lived.
This, in the end, seems to me the power that belongs most distinctively to literary texts. Our encounter with the foregrounded and other defamiliarizing moments in a text oblige us to call on what personal resources we have, so we bring to them our own experiences, concepts, and feelings. In return, the formal processes initiated by the text offer to reshape our ideas, to situate them in a new context with new relationships. Like the reader we have just examined, we take on a new identity for a while and are as a result empowered to see and feel things that in our ordinary lives we would not normally experience.
This suggests that literature is necessary to us as human beings: it confers advantages that are not available from any other medium or activity. And this is why, as I argued earlier, what we need now is not more professional interpretation of literature. We need a better, empirically-based understanding of how literature functions in the lives of ordinary readers. Literary texts, after all, are written for ordinary readers, not for critics. This, I believe, has been largely forgotten in the last fifty years: empirical study will enable us to refocus literary research and education on the distinctive, central functions that literature performs for the individual reader.
Bredin, Hugh. (1996). Onomatopoeia as a figure and a linguistic principle. New Literary History, 27, 555-569.
Brown, R. (1958). Words and things. New York: Free Press.
Coleridge, S. T. (1983). Biographia literaria (W. J. Bate & J. Engell, Eds.). London: Routledge and Kegan Paul. (Original work published 1817)
Eagleton, T. (1983). Literary theory: An introduction. Oxford: Basil Blackwell.
Ellis, J. M. (1997). Literature lost: Social Agendas and the corruption of the humanities. New Haven: Yale University Press.
Fish, S. (1980). Is there a text in this class? The authority of interpretive communities. Cambridge, MA: Harvard University Press.
Fish, S. (1989). Doing what comes naturally: Change rhetoric, and the practice of theory in literary and legal studies. Durham, NC: Duke University Press.
Fish, S. (1995). Professional correctness: Literary studies and political change. Oxford: Clarendon Press.
Gardin, J.-C. (1991). On the way we think and write in the humanities: A computational perspective. In Susan Hockey, Nancy Ide, and Ian Lancashire (Eds.), Research in humanities computing I: Papers from the 1989 ACH-ALLC Conference (pp. 337-345). Oxford: Oxford University Press.
Miall, D. S. (1988). Affect and narrative: A model of response to stories. Poetics, 17, 259-272.
Miall, D. S., & Kuiken D. (1994). Foregrounding, defamiliarization, and affect: Response to literary stories. Poetics 22, 389-407.
O'Grady, W., Dobrovsky, M., & Aronoff, M. (1989). Contemporary linguistics: An introduction. New York: St. Martin's Press.
Pinker, S., & Birdsong, D. (1979). Speakers' sensitivity to rules of frozen word order. Journal of Verbal Learning and Verbal Behavior, 18, 497-508.
Sapir, E. (1929). A study in phonetic symbolism. Journal of Experimental Psychology, 12, 225-239.
Saussure, F. de (1974). Course in general linguistics (W. Baskin trans.). London: Fontana/Collins.
Simon, H. (1994). Literary criticism: A cognitive approach. Stanford Humanities Review, Special Supplement, 4, 1-26. [complete text]